The Internet of Medical Things IoMT

Healthcare Transformation 1st Edition

R. J. Hemalatha
Visit to download the full and correct content document:
https://ebookmeta.com/product/the-internet-of-medical-things-iomt-healthcare-transfor
mation-1st-edition-r-j-hemalatha/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Internet of Medical Things Remote Healthcare Systems

and Applications 1st Edition D. Jude Hemanth

https://ebookmeta.com/product/internet-of-medical-things-remote-
healthcare-systems-and-applications-1st-edition-d-jude-hemanth/

Internet Computing and Internet of Things 1st Edition

Hamid R. Arabnia

https://ebookmeta.com/product/internet-computing-and-internet-of-
things-1st-edition-hamid-r-arabnia/

Intelligent Internet of Things for Smart Healthcare

Systems Durgesh Srivastava

https://ebookmeta.com/product/intelligent-internet-of-things-for-
smart-healthcare-systems-durgesh-srivastava/

Intelligent Internet of Things for Healthcare and

Industry 1st Edition Uttam Ghosh

https://ebookmeta.com/product/intelligent-internet-of-things-for-
healthcare-and-industry-1st-edition-uttam-ghosh/
Internet of Things and Data Mining for Modern
Engineering and Healthcare Applications (Chapman &
Hall/CRC Internet of Things) 1st Edition Ankan
Bhattacharya
https://ebookmeta.com/product/internet-of-things-and-data-mining-
for-modern-engineering-and-healthcare-applications-chapman-hall-
crc-internet-of-things-1st-edition-ankan-bhattacharya/

Internet of Things: Energy, Industry, and Healthcare

1st Edition Arun Kumar Rana (Editor)

https://ebookmeta.com/product/internet-of-things-energy-industry-
and-healthcare-1st-edition-arun-kumar-rana-editor/

Medical Internet of Things: Techniques, Practices and

Applications 1st Edition Anirban Mitra

https://ebookmeta.com/product/medical-internet-of-things-
techniques-practices-and-applications-1st-edition-anirban-mitra/

Emerging Technologies for Healthcare Internet of Things

and Deep Learning Models 1st Edition Monika Mangla
(Editor)

https://ebookmeta.com/product/emerging-technologies-for-
healthcare-internet-of-things-and-deep-learning-models-1st-
edition-monika-mangla-editor/

Healthcare Systems and Health Informatics: Using

Internet of Things (Innovations in Health Informatics
and Healthcare) 1st Edition Pawan Singh Mehra (Editor)

https://ebookmeta.com/product/healthcare-systems-and-health-
informatics-using-internet-of-things-innovations-in-health-
informatics-and-healthcare-1st-edition-pawan-singh-mehra-editor/
Table of Contents
Cover
Title Page
Copyright
Preface
1 In Silico Molecular Modeling and Docking Analysis in Lung
Cancer Cell Proteins
1.1 Introduction
1.2 Methodology
1.3 Results and Discussion
1.4 Conclusion
References
2 Medical Data Classification in Cloud Computing Using Soft
Computing With Voting Classifier: A Review
2.1 Introduction
2.2 Access Control–Based Security
2.3 System Model
2.4 Data Classification
2.5 Related Work
2.6 Conclusion
References
3 Research Challenges in Pre-Copy Virtual Machine Migration in
Cloud Environment
3.1 Introduction
3.2 Existing Technology and Its Review
3.3 Research Design
3.4 Results
 3.5 Discussion
3.6 Conclusion
References
4 Estimation and Analysis of Prediction Rate of Pre-Trained Deep
Learning Network in Classification of Brain Tumor MRI Images
4.1 Introduction
4.2 Classes of Brain Tumors
4.3 Literature Survey
4.4 Methodology
4.5 Conclusion
References
5 An Intelligent Healthcare Monitoring System for Coma Patients
5.1 Introduction
5.2 Related Works
5.3 Materials and Methods
5.4 Results and Discussion
5.5 Conclusion
References
6 Deep Learning Interpretation of Biomedical Data
6.1 Introduction
6.2 Deep Learning Models
6.3 Interpretation of Deep Learning With Biomedical Data
6.4 Conclusion
References
7 Evolution of Electronic Health Records
7.1 Introduction
7.2 Traditional Paper Method
7.3 IoMT
7.4 Telemedicine and IoMT
 7.5 Cyber Security
7.6 Materials and Methods
7.7 Literature Review
7.8 Applications of Electronic Health Records
7.9 Results and Discussion
7.10 Challenges Ahead
7.11 Conclusion
References
8 Architecture of IoMT in Healthcare
8.1 Introduction
8.2 Preferences of the Internet of Things
8.3 IoMT Progress in COVID-19 Situations: Presentation
8.4 Major Applications of IoMT
References
9 Performance Assessment of IoMT Services and Protocols
9.1 Introduction
9.2 IoMT Architecture and Platform
9.3 Types of Protocols
9.4 Testing Process in IoMT
9.5 Issues and Challenges
9.6 Conclusion
References
10 Performance Evaluation of Wearable IoT-Enabled Mesh
Network for Rural Health Monitoring
10.1 Introduction
10.2 Proposed System Framework
10.3 Experimental Evaluation
10.4 Performance Evaluation
10.5 Conclusion
 References
11 Management of Diabetes Mellitus (DM) for Children and Adults
Based on Internet of Things (IoT)
11.1 Introduction
11.2 Materials and Methods
11.3 Results and Discussion
11.4 Summary
11.5 Conclusion
References
12 Wearable Health Monitoring Systems Using IoMT
12.1 Introduction
12.2 IoMT in Developing Wearable Health Surveillance System
12.3 Vital Parameters That Can Be Monitored Using Wearable
Devices
12.4 Challenges Faced in Customizing Wearable Devices
12.5 Conclusion
References
13 Future of Healthcare: Biomedical Big Data Analysis and IoMT
13.1 Introduction
13.2 Big Data and IoT in the Healthcare Industry
13.3 Biomedical Big Data Types
13.4 Biomedical Data Acquisition Using IoT
13.5 Biomedical Data Management Using IoT
13.6 Impact of Big Data and IoMT in Healthcare
13.7 Discussions and Conclusions
References
14 Medical Data Security Using Blockchain With Soft Computing
Techniques: A Review
14.1 Introduction
14.2 Blockchain
 14.3 Blockchain as a Decentralized Security Framework
14.4 Existing Healthcare Data Predictive Analytics Using Soft
Computing Techniques in Data Science
14.5 Literature Review: Medical Data Security in Cloud
Storage
14.6 Conclusion
References
15 Electronic Health Records: A Transitional View
15.1 Introduction
15.2 Ancient Medical Record, 1600 BC
15.3 Greek Medical Record
15.4 Islamic Medical Record
15.5 European Civilization
15.6 Swedish Health Record System
15.7 French and German Contributions
15.8 American Descriptions
15.9 Beginning of Electronic Health Recording
15.10 Conclusion
References
Index
End User License Agreement

List of Illustrations
Chapter 1
Figure 1.1 SOPMA plots for (a) EGFR, (b) K-ras oncogene
protein, and (c) TP53.
Figure 1.2 Ramachandran plots for (a) EGFR, (b) K-ras
oncogene protein, and (c) ...
Figure 1.3 ERRAT plots for (a) EGFR, (b) K-ras oncogene
protein, and (c) TP53.
 Figure 1.4 Verify 3D plots for (a) EGFR, (b) K-ras oncogene
protein, and (c) TP5...
Chapter 2
Figure 2.1 Architecture for PHR system.
Figure 2.2 Data classification in cloud computing. company
information and resou...
Figure 2.3 Fuzzy classification block diagram.
Figure 2.4 Analysis framework architecture.
Chapter 3
Figure 3.1 Cloud computing.
Figure 3.2 Types of cloud.
Figure 3.3 Virtualization.
Figure 3.4 Phases of virtual machine migration.
Figure 3.5 Pre-copy algorithm.
Figure 3.6 Timeline for pre-copy.
Figure 3.7 Improved pre-copy live migration.
Figure 3.8 Iteration of time series–based pre-copy live
migration.
Figure 3.9 Structure and operations of MPLM.
Figure 3.10 Structure of TPO.
Figure 3.11 Structure of multiphase pre-copy live migration.
Figure 3.12 Pre-copy vs. improved pre-copy.
Figure 3.13 Low dirty and high dirty pages based on time
series.
Figure 3.14 Pre-copy vs. TPO.
Figure 3.15 TPO vs. multiphase.
Chapter 4
 Figure 4.1 Design flow of the proposed system.
Figure 4.2 Comparative analysis of pre-trained networks for
brain tumor classifi...
Chapter 5
Figure 5.1 Block diagram.
Figure 5.2 Pulse oximeter and heart rate sensor.
Figure 5.3 Temperature sensor.
Figure 5.4 Complete hardware of the coma patient monitoring
system.
Figure 5.5 Eye blink detection.
Figure 5.6 Yawning detection.
Figure 5.7 Alert message page.
Figure 5.8 Subject testing.
Figure 5.9 Results of heart rate and SpO2.
Figure 5.10 EEG patterns in coma patient.
Chapter 6
Figure 6.1 Basic structural representation of deep learning
process.
Figure 6.2 Simple architecture with hidden layers.
Figure 6.3 Architectural models of deep learning.
Figure 6.4 Architecture of recurrent neural networks.
Figure 6.5 LSTM memory cell.
Figure 6.6 GRU cell.
Figure 6.7 Basic structural framework of convolutional neural
network (CNN).
Figure 6.8 Architectural framework for deep belief networks.
Figure 6.9 Simple architecture of deep stacking networks.
Chapter 9
Figure 9.1 IoMT healthcare system.
Figure 9.2 Platform architecture.
Figure 9.3 Communication protocol overview.
Figure 9.4 The MQTT publish and subscribe model for IoT.
Figure 9.5 CoAP message model.
Figure 9.6 CoAP request/response model.
Figure 9.7 AMQP interaction model with middleware.
Figure 9.8 AMQP capabilities.
Figure 9.9 AMQP for cloud connection.
Figure 9.10 DDS protocol Architecture.
Chapter 10
Figure 10.1 Architecture of wearable IoT-enabled rural health
monitoring system.
Figure 10.2 Body sensor node and its internal architecture.
Figure 10.3 System framework of health monitoring center
(HMC).
Figure 10.4 Sequence diagrams of mesh peering and routing
medical data.
Figure 10.5 GUI alert when the patient’s blood pressure and
sugar is critically ...
Figure 10.6 Energy consumption in HMC.
Figure 10.7 Survival rate.
Figure 10.8 End-to-end delay.
Chapter 11
Figure 11.1 Block diagram of the proposed system.
Figure 11.2 Components of the noninvasive glucose
monitoring system.
 Figure 11.3 Prototype of the glucose monitoring system.
Figure 11.4 Output of glucose monitoring.
Chapter 13
Figure 13.1 Three important V of big data.
Figure 13.2 Applications of big data and IoT in healthcare.
Figure 13.3 Basic details available in an electronic health
record.
Figure 13.4 Commonly used big data management tools.
Chapter 14
Figure 14.1 General architecture and workflow of the
proposed system [7].
Figure 14.2 Remote patient monitoring [8].
Figure 14.3 Blockchain architecture categories [7].
Figure 14.4 Nodes in public vs. private Blockchain [8].
Figure 14.5 Scenarios of using Blockchain in different
healthcare situations [8]...
Figure 14.6 Potential applications of the Blockchain [10].
Figure 14.7 Characteristics of Blockchain.
Chapter 15
Figure 15.1 Evolution of EHR.

List of Tables
Chapter 1
Table 1.1 Physiochemical characters of EGFR, K-ras, and TP53
proteins as determi...
Table 1.2 The number disulfide bonds were quantitated by
Cys_Rec prediction prog...
 Table 1.3 Secondary structure of the EGFR, K-ras oncogene
protein, and TP53.
Table 1.4 Composition of α-helix EGFR, K-ras oncogene
protein, and TP53.
Table 1.5 Validation of the EGFR, K-ras oncogene protein, and
TP53.
Table 1.6 Predicted active sites of the EGFR, K-ras oncogene
protein, and TP53.
Table 1.7 Docking result of the EGFR, K-ras oncogene protein,
and TP53.
Chapter 3
Table 3.1 Type 1 VMM Approach.
Table 3.2 Type 2 VMM Approach.
Chapter 4
Table 4.1 Various symptoms of brain tumors.
Table 4.2 Sample images used for classification purpose.
Table 4.3 Performance of AlexNet pre-trained network.
Table 4.4 Performance of GoogleNet pre-trained network.
Table 4.5 Performance of ResNet101 pre-trained network.
Table 4.6 Comparison of performance metrics between
AlexNet, GoogleNet, and ResN...
Table 4.7 Evaluation of accuracy and processing time of pre-
trained networks.
Chapter 6
Table 6.1 Applications of deep learning networks.
Chapter 10
Table 10.1 Blood glucose classification.
Table 10.2 Blood pressure classification.
Table 10.3 Symptoms and signs of diabetic types.
Table 10.4 DE parameter settings.
Scrivener Publishing
100 Cummings Center, Suite 541J
Beverly, MA 01915-6106
Advances in Learning Analytics for Intelligent Cloud-IoT Systems
Series Editors: Dr. Souvik Pal and Dr. Dac-Nhuong Le
Scope: The role of adaptation, learning analytics, computational
Intelligence, and data analytics in the field of Cloud-IoT Systems is
becoming increasingly essential and intertwined. The capability of an
intelligent system depends on various self-decision making algorithms
in IoT Devices. IoT based smart systems generate a large amount of
data (big data) that cannot be processed by traditional data processing
algorithms and applications. Hence, this book series involves different
computational methods incorporated within the system with the help
of Analytics Reasoning and Sense-making in Big Data, which is centered
in the Cloud and IoT-enabled environments.
The series seeks volumes that are empirical studies, theoretical and
numerical analysis, and novel research findings. The series encourages
cross-fertilization of highlighting research and knowledge of Data
Analytics, Machine Learning, Data Science, and IoT sustainable
developments.
Please send proposals to:
Dr. Souvik Pal
Department of Computer Science and Engineering
Global Institute of Management and Technology
Krishna Nagar
West Bengal, India
souvikpal22@gmail.com
Dr. Dac-Nhuong Le
Faculty of Information Technology, Haiphong University, Haiphong,
Vietnam
huongld@hus.edu.vn
Publishers at Scrivener
Martin Scrivener (martin@scrivenerpublishing.com)
Phillip Carmical (pcarmical@scrivenerpublishing.com)
The Internet of Medical
Things (IoMT)

Healthcare Transformation
Edited by

R.J. Hemalatha

D. Akila

D. Balaganesh
and

Anand Paul
This edition first published 2022 by John Wiley & Sons, Inc., 111 River Street,
Hoboken, NJ 07030, USA and Scrivener Publishing LLC, 100 Cummings Center, Suite
541J, Beverly, MA 01915, USA © 2022 Scrivener Publishing LLC
For more information about Scrivener publications please visit
www.scrivenerpublishing.com.
All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, except as permitted by law.
Advice on how to obtain permission to reuse material from this title is available at
http://www.wiley.com/go/permissions.
Wiley Global Headquarters
111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information
about Wiley products visit us at www.wiley.com.
Limit of Liability/Disclaimer of Warranty
While the publisher and authors have used their best efforts in preparing this work,
they make no representations or warranties with respect to the accuracy or
completeness of the contents of this work and specifically disclaim all warranties,
including without limitation any implied warranties of merchantability or fitness for
a particular purpose. No warranty may be created or extended by sales
representatives, written sales materials, or promotional statements for this work.
The fact that an organization, website, or product is referred to in this work as a
citation and/or potential source of further information does not mean that the
publisher and authors endorse the information or services the organization, website,
or product may provide or recommendations it may make. This work is sold with
the understanding that the publisher is not engaged in rendering professional
services. The advice and strategies contained herein may not be suitable for your
situation. You should consult with a specialist where appropriate. Neither the
publisher nor authors shall be liable for any loss of profit or any other commercial
damages, including but not limited to special, incidental, consequential, or other
damages. Further, readers should be aware that websites listed in this work may have
changed or disappeared between when this work was written and when it is read.
Library of Congress Cataloging-in-Publication Data
ISBN 978-1-119-76883-8
Cover image: Pixabay.Com
Cover design by Russell Richardson
Set in size of 11pt and Minion Pro by Manila Typesetting Company, Makati,
Philippines
Printed in the USA
10 9 8 7 6 5 4 3 2 1
Preface
It is a pleasure for us to put forth this book, The Internet of Medical
Things (IoMT): Healthcare Transformation. Digital technologies have
come into effect in various sectors of our daily lives and it has been
successful in influencing and conceptualizing our day-to-day activities.
The Internet of Medical Things is one such discipline which seeks a lot
of interest as it combines various medical devices and allows these
devices to have a conversation among themselves over a network to
form a connection of advanced smart devices. This book helps to know
about IoMT in the health care sector that involves the latest
technological implementation in diagnostic level as well as therapeutic
level. The security and privacy of maintaining the health records is a
major concern and several solutions for the same has been discussed in
this book. It provides significant advantages for the wellbeing of people
by increasing the quality of life and reducing medical expenses. IoMT
plays a major role in maintaining smart healthcare system as the
security and privacy of the health records further leads to help the
health care sector to be more secure and reliable. Artificial Intelligence
is the other enabling technology that helps IoMT in building smart
defensive mechanisms for a variety of applications like providing
assistance for doctors in almost every area of their proficiencies such as
clinical decision-making. Through Machine Learning and Deep
Learning techniques, the system can learn normal and abnormal
decisions using the data generated by the health worker/professionals
and the patient feedback. This book demonstrates the connectivity
between medical devices and sensors is streamlining clinical workflow
management and leading to an overall improvement in patient care,
both inside care facility walls and in remote locations. This book would
be a good collection of state-of-the-art approaches for applications of
IoMT in various health care sectors. It will be very beneficial for the
new researchers and practitioners working in the field to quickly know
the best methods for IoMT.
• Chapter 1 concentrates on the study of the three-dimensional (3-
D) models of lung cancer cell line proteins (epidermal growth
factor (EGFR), K-Ras oncogene protein and tumor suppressor
(TP53)). The generation and their binding affinities with
curcumins, ellagic acid and quercetin through local docking were
assessed.
• Chapter 2 focuses on cloud computing and electronic health
record system service EHR used to protect the confidentiality of
patient sensitive information and must be encrypted before
outsourcing information. This chapter focuses on the effective use
of cloud data such as search keywords and data sharing and the
challenging problem associated with the concept of soft
computing.
• Chapter 3 elucidates the study of cloud computing concepts,
security concerns in clouds and data centers, live migration and its
importance for cloud computing, and the role of virtual machine
(VM) migration in cloud computing. It provides a holistic approach
towards the pre-copy migration technique thereby explore the way
for reducing the downtime and migration time. This chapter
compares different pre-copy algorithms and evaluates its
parameters for providing a better solution.
• Chapter 4 concentrates on Deep Learning that has gained more
interest in various fields like image classification, self-driven cars,
natural language processing and healthcare applications. The
chapter focuses on solving the complex problems in a more
effective and efficient manner. It elaborates for the reader how
deep learning techniques are useful for predicting and
classification of the brain tumor cells. Datasets are trained using
pre-trained neural networks such as Alexnet, Googlenet and
Resnet 101 and performance of these networks are analysed in
detail. Resnet 101 networks have achieved highest accuracy.
• Chapter 5 illustrates an intelligent healthcare monitoring system
for coma patients that examines the coma patient's vital signs on a
continuous basis, detects the movement happening in the patient,
and updates the information to the doctor and central station
through IoMT. Consistent tracking and observation of these health
issues improves medical assurance and allows for tracking coma
events.
• Chapter 6 details the Deep Learning process that resembles the
human functions in processing and defining patterns used for
decision-making. Deep learning algorithms are mainly designed
and developed using neural networks performing unsupervised
data that are unstructured. Biomedical data possess time and
frequency domain features for analysis and classification. Thus,
deep learning algorithms are used for interpretation and
classification of biomedical big data.
• Chapter 7 discusses how the electronic health records automates
and streamlines the clinician’s workflow and makes the process
easy. It has the ability to generate the complete history of the
patient and also help in assisting for the further treatment which
helps in the recovery of the patient in a more effective way. The
electronic health records are designed according to the
convenience depending on the sector it is being implemented. The
main aim of electronic health records was to make it available to
the concerned person wherever they are, to reduce the work load
to maintain clinical book records and use the details for research
purposes with the concerned persons acknowledgement.
• Chapter 8 elaborates technical architecture of IoMT in relation to
biomedical applications. These ideologies are widely used to
educate people regarding the medical applications using IoMT. It
also gives a detailed study about the future scope of IoMT in
healthcare.
• Chapter 9 provides knowledge on the different performance
assessment techniques and types of protocols that suits best data
transfer and increases safety. The chapter provides the best
protocol which helps in saving energy and is useful for the
customer. It will help the researchers to select the best IoT
protocol for healthcare applications. Testing tools and frameworks
provide knowledge to assess the protocols.
• Chapter 10 addresses the issue of a Health Monitoring Centre
(HMC) in rural areas. The HMC monitors and records continuously
the physiological parameters of the patients in care using wearable
biosensors. The elderly suffering from chronic diseases is
monitored periodically or continuously under the care of the
physician. To enhance the performance of the system a smart and
intelligent mesh backbone is integrated for fast transmission of the
critical medical data to a remote health IOT cloud server.
• Chapter 11 concentrates on Diabetes Mellitus (DM) which is one
of the most widely recognized perilous illnesses for all age groups
in the world. The patients need to settle on the best-individualized
choices about day-by-day management of their diabetes.
Noninvasive glucose sensor used to find out the glucose value of
patients from its fingertip and other sensors also connected to the
patient to get relevant data. A completely useful IoT-based eHealth
stage that wires humanoid robot help with diabetes and planned
successfully. The created platform encourages a constant coupled
network among patients and their caretakers over physical
separation and, in this manner, improving patient’s commitment
with their caretakers while limiting the cost, time, and exertion of
the conventional occasional clinic visits.
• Chapter 12 explores the concepts of wearable health monitoring
systems using IoMT technology. Additionally, this chapter also
provides a brief review about challenges and applications of
customized wearable healthcare system that are trending these
days. The basic idea is to have a detailed study about the recent
developments in IoMT technologies and the drawbacks, as well as
future advancements related to it. The recent innovations,
implications and key issues are discussed in the context of the
framework.
• Chapter 13 provides knowledge on biomedical big data analysis
which plays a huge impact in personalized medicine. Some
challenges in big data analysis like data acquisition, data accuracy,
data security are discussed. Huge volume of data in healthcare can
be managed by integrating biomedical data management. This
chapter will provide brief information on different software that
 are used to manage data in healthcare domain. Impact of big data
and IoMT in healthcare will enhance data analytics research.
• Chapter 14 concentrates on blockchain which is a highly secure
and decentralized networking platform of multiple computers
called nodes. Predictive analysis, soft computing (SC) and
optimization and data science is becoming increasingly important.
In this chapter, the authors investigate privacy issues around large
cloud medical data in the remote cloud. Their proposed framework
ensures data privacy, integrity, and access control over the shared
data with better efficiency. It reduces the turnaround time for data
sharing, improves the decision-making process, and reduces the
overall cost while providing better security of electronic medical
records.
• Chapter 15 discusses the evolution of electronic health record
starting with the history and evolution of the health record system
in the Egyptian era when the first health record was written, all the
way to the modern computerized health record system. This
chapter also includes various documentation procedures for the
health records that were followed from the ancient times and by
other civilizations around the world.

We thank the chapter authors most profusely for their contributors

written during the pandemic.
R. J. Hemalatha
D. Akila
D. Balaganesh
Anand Paul
January 2022
1
In Silico Molecular Modeling and Docking Analysis in
Lung Cancer Cell Proteins
Manisha Sritharan1 and Asita Elengoe2*
1Department of Science and Biotechnology, Faculty of Engineering and Life Sciences, University of
Selangor, Bestari Jaya, Selangor, Malaysia
2Department of Biotechnology, Faculty of Science, Lincoln University College, Petaling Jaya, Selangor,
Malaysia

Abstract
In this study, the three-dimensional (3D) models of lung cancer cell line proteins [epidermal growth
factor (EGFR), K-ras oncogene protein, and tumor suppressor (TP53)] were generated and their
binding affinities with curcumin, ellagic acid, and quercetin through local docking were assessed.
Firstly, Swiss model was used to build lung cancer cell line proteins and then visualized by the PyMol
software. Next, ExPASy ProtParam Proteomics server was used to evaluate the physical and chemical
parameters of the protein structures. Furthermore, the protein models were validated using
PROCHECK, ProQ, ERRAT, and Verify3D programs. Lastly, the protein models were docked with
curcumin, ellagic acid, and quercetin by using BSP-Slim server. All three protein models were
adequate and in exceptional standard. The curcumin showed binding energy with EGFR, K-ras
oncogene protein, and TP53 at 5.320, 2.730, and 1.633, kcal/mol, respectively. Besides that, the ellagic
acid showed binding energy of EGFR, K-ras oncogene protein, and TP53 at –2.892, 0.921, and 0.054
kcal/mol, respectively. Moreover, the quercetin showed binding energy of EGFR, K-ras oncogene
protein, and TP53 at –1.249, –1.154, and –0.809 kcal/mol, respectively. The EGFR had the strongest
bond with ellagic acid while K-ras oncogene protein and TP53 had the strongest interaction with
quercetin. In order to identify the appropriate function, all these potential drug candidates can be
further assessed through laboratory experiments.
Keywords: EGFR, K-ras, TP53, curcumin, ellagic acid, quercetin, docking

1.1 Introduction
Lung cancer is known to be the number one cause of cancer deaths among all the cancer in both men and
women in worldwide. According to a World Health Organization (WHO) survey, lung cancer caused 19.1
deaths per 100,000 in Malaysia, or 4,088 deaths per year (3.22% of all deaths) [1]. Moreover, there was a
record of 1.69 million of deaths worldwide in 2015 due to lung cancer. Furthermore, a research in UK
estimated that there will be 23.6 million of new cases of cancer worldwide each year by 2030 [1]. The
main cause of lung cancer deaths is smoking. Almost 8% of people died because of it. Furthermore, the
second reason is exposure to secondhand smoke. Thus, it is very clear that smoking is the leading risk
factor for lung cancer. However, not everyone who got lung cancer is smokers as many people with lung
cancer are former smokers while many others never smoked at all. Moreover, radiation exposure,
unhealthy lifestyle, secondhand smoke, pollution of air, genetic markers, prolongs inhalation of asbestos,
and chemicals as well as other factors can cause lung cancer non-smokers [2].
Furthermore, it seems that most lung cancer signs do not appear until the cancer has spread, although
some people with early lung cancer do have symptoms. Generally, the symptoms of lung cancer are a
cough that does not go away and instead gets worse, shortness of breath, chest pain, feeling tired or
weak, new onset of wheezing, and some lung cancer can even cause syndrome [3]. On top of that, a
number of tests can be conducted in order to look for cancerous cell such as X-ray image of lung that
could disclose the abnormal mass or nodule, a CT scan to exhibit small lesions in the lungs which may
not detected on X-ray, blood investigations, sputum cytology, and tissue biopsy [4]. Lung cancer
treatments being carried out are adjuvant therapy which may include radiation, chemotherapy, targeted
therapy, or immunotherapy.
Since they originate from the bronchi within the lungs, small-cell lung carcinoma (SCLC) and non-small-
cell lung carcinoma (NSCLC) are the two main clinic pathological classes for lung cancer. They are also
known as bronchogenic carcinomas because they arise from the bronchi within the lungs [4]. Lung
cancer is believed to be arising after a series of continuous pathological changes (preneoplastic lesions)
which very often discovered accompanying lung cancers as well as in the respiratory mucosa of smokers.
Apart from that, the genes involved in lung cancer are epidermal growth factor receptor (EGFR), KRAS,
MET, LKBI, BREF, ALK, RET, and tumor suppressor gene (TP53) [5]. The three most common genes in
lung cancer are EGFR, KRAS, and TP53, and the structure of these genes is explored in thus study. EGFR
is actually transmembrane protein that has cytoplasmic kinase movement and it transduces essential
development factor motioning from the extracellular milieu to the cell. According to da Cunha Santos,
more than 60% of NSCLCs expresses EGFR which has turned into an essential focus for the treatment of
these tumors [6]. In addition, the KRAS mutation is the most widely recognized oncogene driver change
in patients with NSCLC and presents a poor guess in the metastatic setting, making it an imperative focus
for tranquilize advancement. It is difficult to treat patients with KRAS mutations since there is no
targeted therapy yet [7]. Among the mutations, the most common mutation that found to occur in lung
cancer is TP53 mutations and its frequency becomes greater with tobacco consumption [8]. In this study,
three compounds (curcumin, ellagic acid, and quercetin) were used for docking with the three mutant
proteins. Curcumin has excellent safety profile that focus on different infections with solid confirmation
on molecule level. Thus, improvement in formulation criteria can aid in developing therapeutic drug [9].
Next, ellagic acid has the ability to bind with cancer cells to make them inactive and it is also effective to
resist cancer in rats and mice according to a research [10]. Quercetin is a pigment from plant (flavonoid)
which has anti-oxidant and anti-inflammatory effect. It has shown to inhibit the multiplication of cancer
cells according to Pao-Chen Kuo et al. [11].
Bioinformatics is a multidisciplinary discipline that creates methods and software tools for storing,
extracting, organizing, and interpreting biological data. To analyze biological data, a combination of
bioinformatics and computer science, statistics, physics, chemistry, mathematics, and engineering is
useful. Currently, this method is growing rapidly because it is cheap and quite faster than experimental
approaches. Computational biology tools such as protein modeling (e.g., Swiss Model, Easy Modeller, and
Modeller), molecular dynamic simulation (e.g., Gromacs and Amber), and docking (e.g., Autodock version
4.2, AutodockVina, Swissdock, and Haddock) helped design substrate-based drugs to study the
interaction between the target proteins (cancer cell proteins) and ligand (phytocomponents).
The aim of conducting this research is to initiate three-dimensional (3D) models of lung cancer line
proteins (EGFR, K-ras oncogene, and TP53) and to guesstimate their binding affinities with curcumin,
ellagic acid, and quercetin via docking approach.

1.2 Methodology
1.2.1 Sequence of Protein
The entire amino acid sequence of the EGFR (GI: 110002567), K-ras oncogene protein (GI: 186764), and
TP53 (GI: 1233272225) were obtained from National Center for Biotechnology Information Center for
Biotechnology Information (NCBI). Next, EGFR consists of 464 amino acids, K-ras oncogene protein
contains 188 amino acids, while TP53 consists of 346 amino acids.

1.2.2 Homology Modeling

As of now, the 3D models of EGFR, K-ras oncogene protein, and TP53 are not available in Protein Data
Bank (PDB). As a result, the models were started with Swiss Model [12] and then visualized with PyMol
[13].
1.2.3 Physiochemical Characterization
The physical and chemical characters of the protein structures were analyzed by using the ExPASy
ProtParam Proteomics tool [14]. Besides that, hydrophobic and hydrophilic were foreseen by ColorSeq
analysis [15]. Furthermore, The ESBRI program [16] was used to reveal salt bridges in protein
structures, while the Cys Rec program was used to count the number of disulfide bonds [17].

1.2.4 Determination of Secondary Models

The secondary structural properties were discovered using the Alignment Self-Optimized Prediction
Process (SOPMA) [18].

1.2.5 Determination of Stability of Protein Structures

PROCHECK was used to test the protein models [19]. ProQ [20], ERRAT [21], and Verify3D programs
were used to conduct further research [22].

1.2.6 Identification of Active Site

The 3D model of EGFR, K-ras oncogene protein, and TP53 were submitted to active site-prediction server
[23] in order to discover their binding sites.

1.2.7 Preparation of Ligand Model

The tertiary structure of the quercetin, curcumin, and ellagic acid are not openly accessible. The whole
sequence of quercetin, curcumin, and ellagic acid were attained from PubChem, National Center for
Biotechnology Information (2017) [24].

1.2.8 Docking of Target Protein and Phytocompound

The 3D structure of EGFR was docked with quercetin, curcumin, and ellagic acid by using BSP-Slim
server [25]. In addition, the best docking complex model was chosen based on the lowest binding score.
The similar docking method was carried out between the other two protein models and
phytocompounds (quercetin, curcumin, and ellagic acid).

1.3 Results and Discussion

1.3.1 Determination of Physiochemical Characters
The isoelectric point (pI) value quantified for EFGR (pI > 7) specify basic feature while the pI for k-ras
and TP53 (pI < 7) exhibit acidic. Besides that, the molecular weight of EFGR, k-ras oncogene protein, and
TP53 are 50,343.70, 21,470.62, and 38,532.60 Daltons, respectively. The extent of light being by
absorbed by protein at a specific wavelength was used to calculate the extinction coefficient of TYR, TRP,
and CYS residues where for EGFR is 38,305 M/cm, k-ras oncogene protein is 12,170M/cm, and TP53 is
43,025 M/cm. In addition, –R is the negatively (ASP + GLU) and +R is the positively charged (ARG + LYS)
residues in the amino acid sequence. The total of –R and +R for each protein model was described in
Table 1.1. According to the instability index of ExPASy ProtParam, EGFR proteins are classified as stable
because the instability index for both proteins are less than 40 while K-ras oncogene protein and TP53 is
categorized as unstable as the instability index is more than 40. The instability index for EGFR is 35.56,
K-ras oncogene protein is 43.95, and TP53 is 80.17. On top of that, the very low grand average of
hydropathicity (GRAVY) index (a (negative value GRAVY) of EGFR, K-ras oncogene protein, and TP53
denotes their hydrophilic nature (Table 1.1). Apart from that, EFGR, K-ras and TP53 have more polar
residues (41.52%, 53.33%, and 45.29%) than non-polar residues (26.74%, 30.0%, and 27.35%) which
were determined using Color Protein Sequence.
Table 1.1 Physiochemical characters of EGFR, K-ras, and TP53 proteins as determined by ExPASy
ProtParam program.

Protein Length Molecular Pl – +R Extinction Instability Aliphatic GRAVY

weight (kDa) R coefficient index index
EGFR 460 50,343.70 7.10 49 49 38,305 35.56 72.91 –0.269
KRAS 180 21,470.62 8.18 29 31 12,170 43.95 77.18 –0.559
TP53 340 38,532.60 5.64 41 33 43,025 80.17 63.99 –0.592
Furthermore, the structure and function of the protein can be affected by salt bridges. Thus, salt bridge
disruption minimizes the stability of protein [26]. Next, it is also associated with regulation, molecular
recognition, oligomerization, flexibility, domain motions, and thermostability [27, 28].
The greater number of arginine in the protein model enhances the stability of the protein. This is
happens through the electrostatic interactions between their guanidine group [29]. Hence, it was
confirmed that all the protein models are in the identical stable conditions. The outcome of Cys_
Recserver exhibits that the quantity of disulfide bonds in EGFR is 42, K-ras oncogene protein is 5, and
TP53 is 11 (Table 1.2).

1.3.2 Prediction of Secondary Structures

Results from SOPMA analysis shows that random coils dominant among secondary structure
components in the protein models (Figure 1.1). The constitution of alpha helix in EGFR, K-ras oncogene
protein, and TP53 were shown in Table 1.3.
The outcome from this analysis specified that EGFR, K-ras oncogene protein, and TP53 constitutes of 15,
11, and 10α helices, respectively. Besides that, Table 1.4 represents the details of the longest and shortest
alpha helix of all the protein models.

1.3.3 Verification of Stability of Protein Structures

PROCHECK server was used to verify the stereo chemical quality and the geometry of protein models
through Ramachandran plots (Figure 1.2). Furthermore, it was revealed that all the protein structures
are in most favorable region because they had percentage value more than 80% (Table 1.5). Thus, the
standard of these proteins was assessed to be immense and reliable. On top of that, PROCHECK analysis
disclose that a number of residues such as TYR265 and GLU51 for EGFR while LYS180 for K-ras
oncogene protein were located away from energetically favored regions of Ramachandran plot. Besides
that, there are no residues found at forbade region for TP53 protein model.
Thereby, the stereo chemical interpretation of backbone phi/psi dihedral angles deduced that EGFR, K-
ras oncogene protein, and TP53 have low percentage of residues among the protein models. Moreover,
ProQ was utilized in order to validate “the quality” with the usage of Levitt-Gerstein (LG) score and
maximum subarray (MaxSub). All the protein models were within the range for LG and MaxSub score
according to the outcome exhibited for creating a good model (Table 1.5).
Table 1.2 The number disulfide bonds were quantitated by Cys_Rec prediction program.

Protein Cys_Rec Score

 Cys_9 –13.0
EGFR
Cys_13 39.2
Cys_17 100.1
Cys_25 98.3
Cys_26 104.2
Cys_30 104.1
Cys_34 90.5
Cys_42 48.2
Cys_45 56.0
Cys_54 58.2
Cys_58 49.5
Cys_85 55.7
Cys_89 54.9
Cys_101 50.4
Cys_105 44.0
Cys_120 63.0
Cys_123 73.3
Cys_127 75.3
Cys_131 61.6
Cys_156 33.0
Cys_264 45.3
Cys_293 43.8
Cys_300 56.5
Cys_304 46.8
 Cys_309 66.0
Cys_317 65.6
Cys_320 60.2
Cys_329 49.1
Cys_333 42.2
Cys_349 42.2
Cys_352 32.9
Cys_356 62.9
Cys_365 70.2
Cys_373 54.2
Cys_376 54.8
Cys_385 35.8
Cys_389 41.2
Cys_411 78.8
Cys_414 85.1
Cys_418 84.4
Cys_422 26.5
Cys_430 3.7
Cys_12 –28.5
KRAS
Cys_51 –74.2
Cys_80 –72.6
Cys_118 –56.4
Cys_185 –15.2
Cys_124 –19.4
TP53
Cys_135 –1.6
Cys_141 –17.9
Cys_176 –9.1
Cys_182 –45.1
Cys_229 –54.4
Cys_238 1.6
Cys_242 –5.5
Cys_275 –34.4
Cys_277 –51.5
Cys_339 –32.8
ERRAT analysis is used for assessing the protein models which were determined by x-ray
crystallography. Next, the value of ERRAT relies upon the statistics of non-bonded atomic interactions in
the 3D protein structures. The protein is generally accepted as high quality protein if the percentage is
greater than 50%. The ERRAT analysis score result shows that K-ras oncogene protein had the highest at
94.767. Therefore, it can be seen that K-ras oncogene protein has high quality resolution among the
protein models. Besides that, the score value for EGFR is 88.010 while 90.374 for TP53 (Figure 1.3).
The Verify3D server was used to reveal the residues in each protein in which EGFR, K-ras oncogene, and
TP53 had 98.59%, 100.00%, and 92.96% residues, respectively. Next, the average 3D-1D score of all
three proteins are more than 0.2. As a consequence, it specifies that all of the sequences were in line with
its protein model (Figure 1.4). Certainly, the resulting energy minimized EGFR, K-ras oncogene protein,
and TP53 protein models satisfied the standard for evaluation of protein. Hence, the docking analysis
with ligand will be carried out.

Figure 1.1 SOPMA plots for (a) EGFR, (b) K-ras oncogene protein, and (c) TP53.
Table 1.3 Secondary structure of the EGFR, K-ras oncogene protein, and TP53.

Secondary structure Alpha helix (Hh) Extended strand (Ee) Beta turn (Tt) Random coil (CC)
EGFR 16.81 16.81 3.23 64.44
KRAS 43.62 21.81 7.45 27.13
TP53 18.79 18.21 3.18 59.83

Table 1.4 Composition of α-helix EGFR, K-ras oncogene protein, and TP53.

Amino acid Longest alpha helix Residues Shortest alpha helix Number of residues
EGFR α14 14 α3, α6, α11, α15 1
KRAS α11 20 α1,α10 1
TP53 α5 11 α7 1
Figure 1.2 Ramachandran plots for (a) EGFR, (b) K-ras oncogene protein, and (c) TP53.
Table 1.5 Validation of the EGFR, K-ras oncogene protein, and TP53.

Structure Ramachandran plot statistics Goodness factor ProQ

Most Additionally Generously Disallowed Dihedral Covalent Overall LG MaxSub

favored allowed allowed angles forces average score
EGFR 90.6 8.6 0.6 0.3 –0.27 0.02 –0.14 3.814 0.302
KRAS 90.5 8.9 0.0 0.6 –0.12 0.04 –0.04 4.094 0.474
TP53 92.9 7.1 0.0 0.0 –0.16 0.07 –0.06 4.417 0.454
Figure 1.3 ERRAT plots for (a) EGFR, (b) K-ras oncogene protein, and (c) TP53.

1.3.4 Identification of Active Sites

For EGFR, K-ras oncogene protein, and TP53, the BSP-Slim server was used to obtain the active site
protein volume and the residues that form an active site pocket (Table 1.6). The protein volume for
EGFR, K-ras, and TP53 were 837 A3, 718A3, and 647A3, respectively.

1.3.5 Target Protein-Ligand Docking

Based on Murugesan et al. study [30], the plant compounds from methanolic leaf extract of
Vitexnegundoweredocked successfully with cyclooxy-genase-2 (COX-2) enzyme. The phytocompounds
had a better interaction compared with aspirin and ibuprofen. They had a good binding energy and
docking result.
Figure 1.4 Verify 3D plots for (a) EGFR, (b) K-ras oncogene protein, and (c) TP53.
Besides that, four components [1,3-Dioxolane, 2-(3-bromo-5,5,5-trichloro-2,2-dimethylpentyl)-,
Butanoic acid, 2-hydroxy-2-methyl-methyl ester, DL-3,4-Dimethyl-3,4-hexanediol, and Pantolactone]
from Moringaconcanensishad good binding affinity with brain cancer receptors. The binding energies
were –3.90, –2.75, –3.05, and –4.15 kcal/mol. They had the lowest binding energies [31].
According to Deepa et al. study, plant compounds from the ethanolic leaf extract of VitexNegundo [(4S)-2-
Methyl-2-phenylpentane-1,4-diol, 7-Methoxy-2,3-dihydro-2-phenyl-4-quinolone, 3-(tert-
Butoxycarbonyl)-6-(3-benzoylprop-2-yl)phenol, (3R,4S)-4-(methylamino)-1-phenylpent-1-en-3-ol, and
(2S,1’S)-1-Benzyl-2-[1’-(dibenzylamino) ethyl]aziridine] were docked with glucosamine 6 phosphatase
synthase. They had the lowest and most negative value for binding energy (-36.53, –33.57, –35.90, –
33.88, and –37.65 kcal/mol) [32].
Table 1.6 Predicted active sites of the EGFR, K-ras oncogene protein, and TP53.

Protein Volume Residues that forming pocket

EGFR 837 GLY100, ALA 101, ASP102, SER103, TYR104, GLU105, MET106, GLU107, GLU108,
LYS113, LYS115, LYS 116, CYS117, GLU118, GLY119, PRO120, CYS121, ARG122,
LYS123, VAL124, ASN149, THR151, SER152, SER154, THR185, LYS187, GLU188,
THR190, ASN210, GLU212, ILE213, ARG215, LYS242, CYS99
K-ras 718 GLY10, ALA11, CYS12, GLY13, VAL14, GLY15, LYS16, ASP33, PRO34, THR35,
oncogene GLU37, LEU56, ASP57, THR58, ALA59, GLY60, GLN61, GLU62, GLU63, SER65,
protein ARG68, MET72, ALA83, ASN86, LYS88, SER89, GLU91, ASP92, ILE93, HIE94,
HIE95, TYR96, ARG97, GLU98, GLN99, ILE100, ARG102, VAL103
TP53 647 GLU107, ASN109, THR11, PRO128, TYR129, GLN13, GLU130, PRO131, PRO132,
GLU133, VAL134, GLY135, SER136, ASP137, CYS138, THR139, THR140, ILE141,
HIE142, TYR143, TYR16, GLY17, ASN177, SER178, PHE18, ARG19, LEU20, GLY21,
PHE22, LEU23, HIE24, TYR35, ASN40, MET42, THR49, CYS50, PRO51, GLN53,
LEU54, TRP55, VAL56, ASP57, THR59, PRO60, PRO61, THR64
According to Kasilingam and Elengoe study, apigenin successfully docked with p53, caspase-3, and
MADCAM1 using BSP-Slim server. Apigenin was the plant compound while p53, caspase-3, and
MADCAM1 were the target proteins in lung cancer cell line. Apigenin bound strongly with p53, caspase-
3, and MADCAM1 at the lowest binding energies (4.611, 5.750, and 5.307 kcal/mol, respectively) [33].
Based on Ashwini et al. study, coumarin, camptothecin, epigallocate-chin, quercetin, and gallic acid were
screened for potential binding with caspase-3 (target protein) in human cervical cancer cell line (HeLa).
Coumarin had the strongest interaction with caspase-3 at the lowest binding affinity (–378.3 kJ/mol).
Therefore, it could be a potential anti-cancer drug. However, gallic acid had the least interaction with
caspase-3 at the lowest binding energy (–181.3 kJ/mol). The docking approach was carried out using
Hex 8.0.0 docking software [34].
Chakrabarty et al. study demonstrated that 1-hexanol and 1-octen-3-ol suppressed the enzyme activity
of Ach (PDB id: 2CKM) and BACE1 (PDB id: 4IVT). Ach and BACE1 are the proteins responsible for
Alzheimer disease. 1-hexanol and 1-octen-3-ol were the plant compounds derived from leaf extract of
Lantana Camera (L.). Glide Standard Precision (SP) ligand docking was performed to determine the
binding energy. The results show that 1-hexanol and 1-octen-3-ol bound strongly with Ach at –2.291 and
–2.465 kJ/mol, respectively. Whereas, 1-hexanol and 1-octen-3-ol had the lowest binding affinity with
BACE 1 at –0.948 and –1.267 kJ/mol, respectively. 1-octen-3-ol may have the potential to be an effective
drug against Alzhemeir disease. It had the best interaction with both enzymes (Ach and BACE1) when
compared with 1-hexanol [35].
Based on Supramaniam and Elengoe study, glycyrrhizin successfully docked with p53, NF-kB-p105, and
MADCAM1 using BSP-Slim server. Glycyrrhizin was the plant compound while p53, NF-kB-p105, and
MADCAM1 were the target proteins in breast cancer cell line. Glycyrrhizin bound strongly with p53, NF-
kB-p105, and MADCAM1 at the lowest binding affinities (-4.040, –5.127, and –5.251 kcal/mol,
respectively). Therefore, glycyrrhizin could be a potential drug in breast cancer treatment [36].
According to Elengoe and Sebestian study, p53, adenomatous polyposis coli (APC), and EGFR were
generated using homology modeling approach. These proteins were the target proteins. They were
docked successfully with plant compounds such as allicin, epigallocatechin-3-gallate, and gingerol. Plant
compounds were used as ligands in docking process. p53 had the most stable interaction with the allicin
among the three target proteins. p53 docked with allicin at the lowest binding energy of 4.968. However,
the other target proteins had the good docking score too [37].
In this study, EGFR is successfully docked with quercetin, curcumin, and ellagic acid by using the BSP-
Slim server. The same target protein-phytocompound complex docking method was repeated with K-ras
oncogene protein and TP53. Furthermore, the most suitable docking complex was selected based on the
lowest binding energy (DGbind). Results of docking showed that EGFR had a strong bond with ellagic
acid since it was the most favorable with the lowest energy value (–2.892 kcal/mol) when compared to
curcumin and quercetin (Table 1.7). In addition, there was strong interaction between K-ras oncogene
protein and quercetin with lowest energy (–1.154 kcal/mol) that was most favorable when compared to
curcumin and ellagic acid. In addition, the strongest interaction for TP53 was with quercetin when
compared to other two compounds with lowest energy (0.809 kcal/mol) according to the docking
analysis.
Table 1.7 Docking result of the EGFR, K-ras oncogene protein, and TP53.
Protein Compounds Binding energy (kcal/mol)
Curcumin 5.320
EGFR
Ellagic acid –2.892
Quercetin –1.249
K-ras oncogene protein Curcumin
2.730

Ellagic acid 0.921

Quercetin –1.154
 Curcumin 1.633
TP53
Ellagic acid 0.054
Quercetin –0.809

1.4 Conclusion
In a nutshell, EGFR was successfully docked with curcumin, ellagic acid, and quercetin. Besides that, the
same approach of docking simulation was performed for K-ras oncogene protein and TP53. Among the
three protein models, EGFR had a strong interaction with ellagic acid due to the lowest energy value
while K-ras oncogene protein and TP53 had a strong interaction with quercetin as the binding energy
was the lowest. Consequently, result from this study will aid in designing a suitable structure-based drug.
However, wet lab must be carried out to verify the results of this study.

References
1. Cancer Research UK, Worldwide cancer statistics, 2012, https://www.cancerresearchuk.org/health-
professional/cancer-statistics/worldwide-cancer#collapseZero.
2. American Cancer Society, Lung cancer prevention and early detection, 2016,
https://www.cancer.org/cancer/lung-cancer/prevention-and-early-detection/signs-and-
symptoms.html.
3. American Cancer Society, Causes, risk factors and prevention, 2016,
https://www.cancer.org/cancer/non-small-cell-lung-cancer/causes-risks-prevention/ what causes.
4. Mayo Clinic, Lung cancer, 2018, https://www.mayoclinic.org/diseases-conditions/lung-
cancer/diagnosis-treatment/drc-20374627.
5. El-Telbany, A. and Patrick, C.M., Cancer genes in lung cancer. Genes Cancer, M3, 7–8, 467–480, 2012.
6. Santos, D.C., Sheperd, F.A., Tsao, M.S., EGFR mutations and lung cancer. Annu. Rev. Pathol.: Mechanisms
of Disease, 6, 49–69, 2016.
7. Bhattacharya, S., Socinski, M.A., T.F., KRAS mutant lung cancer: progress thus far on an elusive
therapeutic target. Clin. Transl. Med., 4, 35, 2015.
8. Halverson, A.R., Silwal-Pandit, L., Meza-Zepeda, L.A. et al., TP53 mutation spectrum in smokers and
never smoking lung cancer patients. Front. Genet., 7, 85, 2016.
9. Basnet, P. and Skalko-Basnet, N., Curcumin: An anti-inflammatory molecule from a curry spice on the
path to cancer treatment. Molecules, 6, 6, 4567–4598, 2011.
10. Healthline, Whyellagic acid is important?, 2020, https://www.healthline.com/health/ellagic-acid.
11. Kuo, P.-C., Liu, H.-F., Chao, J.-I., Survivin and p53 modulate quercetin-induced cell growth inhibition
and apoptosis in human lung carcinoma cells. J. Biol. Chem., 279, 53, 55875–55885, 2004.
12. Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T. et al., SWISS-MODEL:
Modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res.,
42, W252-8, 2014.
13. Delano, W.L., The PyMOL molecular graphics system, 2001, http://www.pymol.org.
14. Gasteiger, E., Hoogland, C., Gattiker, A. et al., Protein identification and analysis tools on the ExPASy
server, in: The proteomics protocols handbook, J.M. Walker (Ed.), Humana Press, Totowa, 2015.
15. Prabi, L.G., Color protein sequence analysis, 1998, https://npsaprabi.ibcp.fr/cgi-
bin/npsa_automat.pl?page=/NPSA/npsa_color.html.
16. Costantini, S., Colonna, G., Facchiano, A.M., ESBRI: a web server for evaluating salt bridges in proteins.
Bioinformation, 3, 137–138, 2008.
17. Roy, S., Maheshwari, N., Chauhan, R. et al., Structure prediction and functional characterization of
secondary metabolite proteins of Ocimum. Bioinformation, 6, 8, 315–319, 2011.
18. Geourjon, C. and Deleage, G., SOPMA: significant improvements in protein secondary structure
prediction by consensus prediction from multiple alignments. Comput. Appl. Biosci., 11, 681–684, 1995.
19. Laskowski, R.A., MacArthur, M.W., Moss, D.S. et al., PROCHECK: a program to check the stereo
chemical quality of protein structures. J. Appl. Cryst., 26, 283–291, 1993.
20. Wallner, B. and Elofsson, A., Can correct protein models be identified? Protein Sci., 12, 1073–1086,
2003.
21. Colovos, C. and Yeates, T.O., Verification of protein structures: patterns of non-bonded atomic
interactions. Protein Sci., 2, 1511–1519, 1993.
22. Eisenberg, D., Luthy, R., Bowie, J.U., VERIFY3D: assessment of protein models with three- dimensional
profiles. Methods Enzymol., 77, 396–404, 1977.
23. Jayaram, B., Active site prediction server, 2004, http://www.scftio-iitd.res.in/dock/ActiveSite.jsp.
24. National Center for Biotechnology Information, Pubchem., 2017, https://pubchem.ncbi.nlm.nih.gov/.
25. Hui, S.L. and Yang, Z., BSP-SLIM: A blind low-resolution ligand-protein docking approach using
theoretically predicted protein structures. Proteins, 80, 93–110, 2012.
26. Kumar, S., Tsai, C.J., Ma, B. et al., Contribution of salt bridges toward protein thermo-stability. J.
Biomol. Struct. Dyn., 1, 79–86, 2000.
27. Kumar, S. and Nussinov, R., Salt bridge stability in monomeric proteins. J. Mol. Biol., 293, 1241–1255,
2009.
28. Kumar, S. and Nussinov, R., Relationship between ion pair geometries and electrostatic strengths in
proteins. Biophys. J., 83, 1595–1612, 2002.
29. Parvizpour, S., Shamsir, M.S., Razmara, J. et al., Structural and functional analysis of a novel
psychrophilic b-mannanase from Glaciozyma Antarctica PI12. J. Comput. Aided Mol. Des., 28, 6, 685–698,
2014.
30. Murugesan, D., Ponnusamy, R.D., Gopalan, D.K., Molecular docking study of active phytocompounds
from the methanolic leaf extract of vitexnegundo-against cyclooxygenase-2. Bangladesh J. Pharmacol., 9,
2, 146–53, 2014.
31. Balamurugan, V and Balakrishnan, V., Molecular docking studies of Moringaconcanensisnimmo leaf
phytocompounds for brain cancer. Res. Rev.: J. Life Sci., 8, 1, 26–34, 2018.
32. Santhanakrishnan, D., Sipriya, N., Chandrasekaran, B., Studies on the phytochemistry, spectroscopic
characterization and antibacterial efficacy of Salicornia Brachiata. Int. J. Pharm. Pharm. Sci., 6, 6430–
6432, 2014.
33. Kasilingam, T. and Elengoe, A., In silico molecular modeling and docking of apigenin against the lung
cancer cell proteins. Asian J. Pharm. Clin. Res., 11, 9, 246–252, 2018.
34. Ashwini, S., Varkey, S.P., Shantaram, M., Insilico docking of polyphenolic compounds against Caspase
3-HeLa cell line protein. Int. J. Drug Dev. Res., 9, 28–32, 2017.
35. Chakrabarty, N., Hossain, A., Barua, J., Kar, H., Akther, S., Al Mahabub, A., Majumder, M., Insilico
molecular docking of some isolated selected compounds of lantana camera against Alzheimer’s disease.
Biomed. J. Sci. Tech. Res., 12, 2, 9168–9171, 2018.
36. Supramaniam, G. and Elengoe, A., In silico molecular docking of glycyrrhizin and breast cancer cell
line proteins, in: Plant-Derived Bioactives, 1, 575–589, 2020.
37. Elengoe, A. and Sebestian, E., In silico molecular modelling and docking of allicin, epigallocatechin-3-
gallate and gingerol against colon cancer cell proteins. Asia Pac. J. Mol. Biol. Biotechnol., 4, 2851–67, 2020.
*Corresponding author: asitaelengoe@yahoo.com
2
Medical Data Classification in Cloud
Computing Using Soft Computing
With Voting Classifier: A Review
Saurabh Sharma1*, Harish K. Shakya1† and Ashish Mishra2‡
1Dept. of CSE, Amity School of Engineering & Technology, Amity
University (M.P.), Gwalior, India
2Department of CSE, Gyan Ganga Institute of Technology and
Sciences, Jabalpur, India

Abstract
In the current context, a tele-medical system is the rising medical
service where health professionals can use telecommunication
technology to treat, evaluate, and diagnose a patient. The data in the
healthcare system signifies a set of medical data that is sophisticated
and larger in number (X-ray, fMRI data, scans of the lungs, brain,
etc.). It is impossible to use typical hardware and software to
manage medical data collections. Therefore, a practical approach to
the equilibrium of privacy protection and data exchange is required.
To address these questions, several approaches are established,
most of the studies focusing on only a tiny problem with a single
notion. This review paper analyzes the data protection research
carried out in cloud computing systems and also looks at the major
difficulties that conventional solutions confront. This approach helps
researchers to better address existing issues in protecting the
privacy of medical data in the cloud system.
Keywords: Medical data, soft computing, fuzzy, cloud computing,
data privacy, SVM, FCM
2.1 Introduction
There are many definitions in Electronic Health Record (EHR), such as
the electronic record that holds patient information on a health record
system operated by healthcare providers [1]. Although EHR has a good
effect on healthcare services, development in many healthcare
institutions globally, particularly in poor nations, is delayed due to
numerous common problems. Patient data security has been a problem
since the beginning of medical history and is an important issue in
modern day. Initiated by the idea of confidentiality, the Oath of
Hippocrates has proved to be an honorable activity in clinical and
medical ethics. It is of highest importance to protect the privacy and
confidentiality of patient information; security is trustworthy. Medical
record security generally involves privacy and confidentiality [2]. Cloud
computing provides the option of accessing massive amounts of patient
information in a short period of time. This makes it easier for an
unauthorized person to obtain patient records. It confirms this feeling
by saying “illegal access to traditional medical records (paper-based)
has always been conceivable, but computer introduction increases a
little problem to a large problem.”
Cloud computing is a concept for easy, on-demand access to a common
pool of configurable computer resources (e.g., networks, servers,
storage, applications, and services), which may easily be provided and
disclosed with minimal administration effort or engagement from
service providers [4]. The newest, most exciting, and comprehensive
solution in the world of IT is cloud computing. Its major purpose is to
use the Internet or intranet to exchange resources for users [5]. Cloud
computing is an affordable, automatically scalable, multi-tenant, and
secure cloud service provider platform (CSP).

2.1.1 Security in Medical Big Data Analytics

Big data is complex and uncomplicated by its very nature and requires
suppliers to take a close look at their techniques to collection, storage,
analysis, and presentation of their data to personnel, business partners,
and patients.
What are some of the most challenging tasks for enterprises when
starting up a big data analytics program, and how can they overcome
these problems to reach their clinical and financial goals?

2.1.1.1 Capture
All data comes from someone, but regrettably, it is not always from
someone with flawless data management habits for many healthcare
providers. Collecting clean, comprehensive, precise, and correctly
structured data for numerous systems is a constant battle for
businesses, many of whom are not on the gaining side of the conflict.
In a recent investigation at an ophthalmology clinic, EHR data were
only 23.5% matched by patient-reporting data. When patients reported
three or more eye problems, their EHR data were absolutely not in
agreement.
Poor usability of EHRs, sophisticated processes, and an incomplete
understanding why big data is crucial to properly collect all can
contribute to quality problems that afflict data during its life cycle.
Providers can begin to improve the data capture routines by
prioritizing valuable data types for their specific projects, by enlisting
the data management and integrity expertise of professional health
information managers, and by developing clinical documentation
improvement programs to train clinicians on how to ensure data are
useful for downstream analysis.

2.1.1.2 Cleaning
Health providers are familiar with the necessity of cleanliness in both
the clinic and the operating room, but are not aware of the importance
of cleaning their data.
Dirty data can swiftly ruin a large data analytics project, especially if
multiple data sources are used to capture clinical or operational
elements in slightly different formats. Data cleaning—also known as
cleaning or scrubbing—guarantees accuracy, correctness, consistency,
relevance, and in no way corruption of datasets.
While most data cleaning activities are still done manually, certain IT
vendors provide automated scrubbing instruments that compare,
contrast, and rectify big data sets using logic rules. These technologies
may grow more sophisticated and accurate as machine learning
techniques continue to progress rapidly, lowering time and cost
necessary to guarantee high levels of accuracy and integrity in health
data stores.

2.1.1.3 Storage
Clinicians at the front line rarely worry about the location of their data,
yet it is a critical cost, safety, and performance issue for the IT
department. Due to the exponential growth in the amount of health
data, several suppliers can no longer manage the costs and implications
on local data centers.
While many firms are more convenient to store data in the premises,
which promises control over security, access, and up-time, the on-site
server network can be costly, hard to operate, and prone to data silo
production in various departments.
Cloud storage is becoming more and more common as costs decrease
and reliability increases. Nearly, 90% of healthcare firms use some
cloudbased IT infrastructure, including warehousing and applications
in a 2016 survey.
The cloud promises a smooth recovery from disasters, reduced upfront
costs, and simpler expansion—even though enterprises have to be
exceedingly careful to select partners who understand the significance
of HIPAA and other compliance and safety issues for health.
Many firms have a hybrid approach to their data store initiatives, which
can offer providers with diverse access and storage requirements the
most flexible and workable solution. However, providers should be
careful to ensure that separate systems can communicate and share
data with other sectors of the company when appropriate while
establishing a hybrid infrastructure.
2.1.1.4 Security
Data security for healthcare businesses is the number one issue,
particularly following a fast fire succession of high-profile violations,
hackings, and ransomware outbreaks. From phishing assaults, viruses,
and laptops left accidently in a cab, health information is exposed to an
almost endless range of dangers.
The HIPAA Security Rule offers a broad set of technological guarantees
for PHI storage organizations, including transmission security,
authentication procedures and access, and integrity and auditing
measures.
These precautions really lead to common sense safety processes, such
as the use of up-to-date anti-virus software, the setup of firewalls, the
encryption of sensitive data, and multi-factor authentication.
However, even the most closely secured data center can be overcome by
personnel who tend to give priority over long software updates and
sophisticated limits on their access to data or software.
Health organizations should often remind their staff members of the
important nature of data security standards and continuously examine
who has access to high-value data in order to prevent damage caused
by malevolent parties.

2.1.1.5 Stewardship
Health data has a long shelf-life, especially on the clinical side. In
addition to keeping patient data accessible for at least 6 years,
clinicians may choose to use de-identified datasets for research
projects, which is vital for continued stewardship and cure. For
additional objectives, such as quality measurement or performance
benchmarking, data may also be repurposed or re-assessed.
Understanding when and for what purposes the data were created—as
well as who utilized it previously, why, how, and when—is vital to
academics and data analysts.
The development of complete, accurate, and up-to-date metadata is an
important component of a successful data management plan. Metadata
enables analysts to precisely duplicate earlier questions that are critical
for scientific investigations and proper benchmarking and prevents the
creation of “data trash”.
Health organizations should employ a data manager to produce and
curate valuable metadata. A data controller may ensure that all pieces
have standard definitions and formats, are properly documented from
creation to deletion, and remain valuable for the tasks involved.

2.2 Access Control–Based Security

Access control is a mechanism to ensure that users are who they say
they are and have enough access to company data.
Access control at a high level is a selective restriction of data access. It
comprises two primary components: authentication and authorization,
as explained by Daniel Crowley, IBM’s X-Force Red research manager
with a focus on data security.
Authentication is a technique used to check that someone claims to be.
Authentication alone is not enough to protect data, as noted by Crowley.
What is required is an additional authorization layer that assesses if a
user should be authorized to access or execute the transaction.

2.2.1 Authentication
Authentication is the process of establishing trust in user identity.
Certification assurance levels will be in accordance with the application
and nature and sensitivity to the risk involved. An increasing number of
cloud providers are reached using their previously certified standards
and user support and administration applications and data. Also, a
common two-factor authentication, in the form of strong
authentication, is, for example, to be used as online banking. In theory,
it should be protected using strong authentication networks. The
stricter requirements apply mainly to CSP employees. They also have
access to IT resources; just for example, it will be provided through
strong authentication, using a chip card or USB stick that can be
generated by hardware through hardware-based password
authentication system or media. This is really necessary to use on the
Internet. He went on to establish strict procedures that are the basis of
all relationships of trust between participants for relationships
between two actors. After the trust relationship is established through
a series of trusted from a certification authority, participants can be
used to authenticate each other in connection with [3]. There are a
variety of authentication methods and techniques that organizations
can choose as follows.

2.2.1.1 User Password Authentication

Authentication is the process of identifying users who ask for system,
network or device access. Access control frequently determines user
identity using credentials such as login and password. Additional
authentication technologies, such as biometric and authentication
applications, are also utilized to authenticate user identification.

2.2.1.2 Windows-Based User Authentication

Typically, the list is stored in the Windows Active Directory for the
organization. The access control framework must be enabled to provide
authentication for the user’s primary domain controller (PDC).

2.2.1.3 Directory-Based Authentication

To continue our expansion in business volume, often millions of users
trying to use resources simultaneously. In such a scenario, the
authentication body should be able to provide faster authentication. A
directory-based authentication technique that is used to respond goes
to the store LDAP user directory to verify user credentials.

2.2.1.4 Certificate-Based Authentication

It is also the user where you can connect digital ID, strong
authentication technology. It released the authority for digital ID
verification, also known as a digital ID trustworthy digital certificate. To
ensure identification, a user has checked a variety of other parameters.

2.2.1.5 Smart Card–Based Authentication

This certificate is used as a second factor [13]. Smart card is the
smallest co-processor data operation cryptographic tool.
2.2.1.6 Biometrics
This is a strong certification [9]. The third aspect of authentication to be
done is based on the user. He said that those that they know
(username) and (either network or token) or after work that they have
(retinal scan, fingerprint or thermal scanning). In cases necessary for
data, such as military/defense, are confidential.

2.2.1.7 Grid-Based Authentication

It is used as a second authentication factor. The user knew that
(authenticated by the authentication username password), and then
they asked her (grid card information). Entrust Identity Protector
provides this certificate.

2.2.1.8 Knowledge-Based Authentication

In order to gain additional confidence in the identity of those users,
keep in mind that the challenge attacker [2] is unlikely to be able to
provide. On the basis of “shared secret”, the organization questions the
user, when appropriate, to allow user information that has been
through the registration process, or how to go on related to the
confirmation of the previous transaction wants to do.

2.2.1.9 Machine Authentication

Authentication of a machine is the authorization of automated
communication from person-to-machine (M2M) by verification of
digital certificates or digital credentials.
Digital certificates used in machine permits are like a digital passport
that provides a trustworthy identification for secure information
exchange on the Web. Digital credentials are similar to types of ID and
password issued by the machine.
Machine authentication is used to allow machine interactions on cable
and wireless networks in order to allow autonomous interaction and
information sharing between computers and other machines. Machine
authentication operations can be carried out with simple devices such
as sensors and infrastructure meters.
2.2.1.10 One-Time Password (OTP)
A password is generated dynamically and is valid only once. The
advantage of a one-time password is that if an intruder does not hack it,
then he cannot use it anymore. There are two types of OTP generator
traces: synchronous and asynchronous. One-time password (OTP)
systems provide a mechanism for logging on to a network or service
using a unique password that can only be used once, as the name
suggests. The static password is the most common authentication
method and the least secure.

2.2.1.11 Authority
The integrity of cloud computing needs an important information
security to maintain relevant authority. It follows the following controls
and privileges in the process stream in cloud computing. The rights
management system should ensure that each role (including metadata)
can see the need to obtain the data function. Access control should be
based and the established role goes on and officers should be reviewed
regularly. In general, the model of least privilege should be used, and
the user and administrator only have the necessary rights for the CSP to
enable them to achieve their functions [14].

2.2.1.12 Global Authorization

Subscribing to global organizations (as many as access control
decisions) and rules and regulations (such as a limited user) must be
lost locally. The decision should be two pieces of information provided.
Subscribed virtual organizations are using the grid. In the early version
of Globus software, subscription information will be found on the local
network. The network [12] is mapped to the DN Mapfail account in that
they require an account on all of the resources they wish to use. The
authorization process performed on the Grid DAS side exploiting
Community Authorization extensions (VO-based) present into the
user's credentials (e.g., proxy).
Another random document with
no related content on Scribd:
Mississippi river. We can now look across a gorge from the coaches
of the Pennsylvania Railroad, beyond Altoona, and see the grade of
the old Portage Railway.

Fig. 28. Broad Street Station, Philadelphia: Pennsylvania

Railroad
The canal almost put out of business the Conestoga wagons on
the dusty pike which had seen so much travel by way of Carlisle and
Bedford. But the people did not stop with a canal. Like the men of
New York, they wanted something even better than that. They
wished to have a railroad all the way, and in 1846 the Pennsylvania
Railroad Company was incorporated. By this time it was very well
known that railroads were successful both in America and in
England, and that steam was better than horses.
 Over the Allegheny Front a route was found where the grades
were not too steep for locomotives. The grade, of course, had been
the one great hindrance to the whole project, and when this difficulty
was overcome there was no reason why passengers should not be
carried from Philadelphia to Pittsburg, or a load of iron from Pittsburg
to Philadelphia, without changing cars. In the year 1854 the
Pennsylvania people triumphed, for they had conquered the
mountains and could run trains from the banks of the Delaware to
the Ohio river.
If we leave Philadelphia by the great Broad street station of the
Pennsylvania Railroad, we shall pass out among the pleasant homes
of West Philadelphia and through the fine farms of the Pennsylvania
lowlands, until we come, in about an hour and a half, to the staid old
city of Lancaster. We have been here before, to learn of turnpikes
and Conestoga freighters.
The next stop, if we are on an express train, will be at
Harrisburg, a little more than a hundred miles from Philadelphia. We
have now come from the Delaware to the Susquehanna, and are
close to the mountains. Before we go in among them let us see
Harrisburg. It is a city of fifty thousand people, and lies along the
east bank of the Susquehanna, which here is a great river a mile
wide, having gathered its tribute of waters from hundreds of
branching streams in Pennsylvania and New York.

Fig. 29. Bridge, Pennsylvania Railroad, above Harrisburg

 Not far to the east a small stream runs parallel to the main river,
and the larger part of Harrisburg is on higher ground between the
two. On the highest part of this ridge is the state capitol, a great
building but recently finished. Harrisburg is at the right point for the
state government. It is not in the center of the state, to be sure, but it
is at the rear of the lowlands which reach in from the sea, and is just
outside the great gateway where roads from all the northern,
western, and central uplands come out on the plain. It is a
convenient center for coal and iron, and hence one sees along the
river below the city many blast furnaces, rolling mills, and factories.
To the northeast rich, open lands stretch along the base of Blue
mountain, and railroads join Harrisburg to Reading, Allentown,
Bethlehem, and Easton. To the southwest bridges cross the
Susquehanna, and roads run to Carlisle, Hagerstown, and other
cities of the Great Valley (Chapter XI).
Thus the Pennsylvania Railroad, running northwest from
Philadelphia, crosses at Harrisburg other roads that run to the
southwest. As hamlets often gather about “four corners” in the
country, so cities grow up where the great roads of the world cross
each other.

Fig. 30. Pennsylvania

Railroad Shops, Altoona
 Leaving Harrisburg behind, we pass the splendid new bridge of
the Pennsylvania Railroad, across the Susquehanna (Fig. 29), and
go through the gap in Blue mountain. Soon we turn away from the
main river and enter the winding valley of the Juniata. The grades
are easy, the roadbed is smooth, and by deep cuts through the rocks
the curves have been made less abrupt. It is only when one looks
out of the car window that the land is found to be rugged and
mountainous.
All the greater valleys and ridges of the mountain belt of
Pennsylvania run northeast and southwest. The last of these to be
crossed on our journey is Bald Eagle valley, from which the
Allegheny Front rises to the northwest.
In this valley, near the place where the Portage Railway began to
scale the heights, and a little more than a hundred miles from
Pittsburg, the Pennsylvania Railroad Company in 1850 founded a
town and called it Altoona. Here they started shops, which have now
grown to notable importance. The town became a city eighteen
years after it was begun, and has to-day about forty thousand
inhabitants. In the railway shops alone may be found nine thousand
men repairing and building locomotives, passenger coaches, and
freight cars. The Pennsylvania Railroad Company is now founding a
great school in Altoona, where young men may be taught to become
skillful and efficient in railway service.

Fig. 31. Horseshoe Curve, Pennsylvania Railroad

Altoona looks new, and with its endless freight yards, its noisy
shops, and its sooty cover of smoke from burning soft coal, it is very
different from quiet Lancaster, which was old when forests covered
the site of Altoona.
On our way to Pittsburg we are soon pulling up the Allegheny
Front by a great loop, or bend, which enables the tracks to reach the
summit more than a thousand feet above Altoona. Nestling within the
great bend is a reservoir of water to supply the houses and shops of
the city lying below. Passing the highest point, we find ourselves
descending the valley of the Conemaugh river to Johnstown, and
surrounded by the high lands of the Allegheny plateau.
Johnstown is much older than Altoona, for it was settled in 1791,
but it has not grown so fast, and has only about as many inhabitants
as the city of railroad shops. Most people know of Johnstown
because of the flood which ruined the place in 1889. Several miles
above the town was a reservoir more than two miles long and in
several places one hundred feet deep. After the heavy rains of that
spring the dam broke on the last day of May, and the wild rush of
waters destroyed the town. Homes, stores, shops, and mills were
torn away and carried down the river. Clara Barton of the Red Cross,
who went to Johnstown as soon as she could get there, says that the
few houses that were not crushed and strewn along the valley were
turned upside down.
More than two thousand men, women, and children lost their
lives, and those that were left were in mourning and poverty. The
whole land sent in its gifts of money, clothing, and food, and the town
was built up again into a prosperous city. Near the city are found
coal, iron, limestone, and fire clay, and these things make it easy to
establish iron works. The Cambria Steel Company gives work to ten
thousand men in its shops, mines, and furnaces.
The main line of the Pennsylvania Railroad runs down the
rugged Conemaugh valley through Johnstown, and is its chief
means of traffic. As we go on to the west we near Pittsburg, but first
we pass through a number of stirring towns. At one place fire bricks
are made, and the clay for molding them and the coal for burning
them are found in the same hill. In another town there are coal mines
and glass works. Farther west the Pennsylvania road has more
repair shops, and Braddock is the great Carnegie town. We shall see
why many thriving young cities have grown up in this region when
we take up Pittsburg, about which they are all clustered.
At Pittsburg we pull into one of the finest railway stations in the
United States. We may stop in the city of coal and iron, or we may
go on to the west, over one of the main arms of the Pennsylvania
Railroad system. If we take the northern branch, it will carry us
across Ohio to Fort Wayne in Indiana and to Chicago. If we board a
train on the southern arm, we shall go through Columbus and
Indianapolis, and be set down on the farther side of the Mississippi
river at St. Louis.
North and south from the great east and west trunk lines run
many shorter roads, or “spurs.” On the east there is a network of
short roads in New Jersey, and one of the busiest parts of the whole
system is that which joins Washington to Baltimore, Philadelphia,
and New York.
 Fig. 32. Rock Cut, along the Line of the Pennsylvania
Railroad
West from Philadelphia for a long distance there are four tracks,
and on either side may be seen neat hedges, such as one finds
along the railways of England. In the mountains it is often hard to
make a roadbed wide enough for four tracks, and hence there may
be only three or even two in some places. No doubt four will in time
be built through to Pittsburg, for many millions of dollars are spent in
improving the road. Instead of having a long circuit around the hills,
tunnels and vast cuts in the bed rock are made so as to straighten
the line. Thus both passenger and freight trains are able to make
better time, and the road can carry the stores of iron and coal which
are found in the lands on either side.
 Some of the freight yards are always crowded with cars, and at
Harrisburg the company is building separate tracks around the city,
so that through freight trains need not be delayed.
At New York the Pennsylvania Railroad now has its station on
the New Jersey side of the Hudson river, but it is building a tunnel
under the river. The company has already bought several city blocks
and has torn away the buildings. Here it will build one of the greatest
passenger stations in the world. The tunnel will run on to the east,
under the streets and shops of Manhattan, and under the East river.
Thus under New York and its surrounding waters trains can go to the
east end of Long Island.
Pennsylvania has told us the same story that we learned from
New York. We read it again: first, how the Indian’s path was beaten
deeper and wider by the hoofs of the pack horse, bearing goods to
sell and barter in the wilderness; then how strips of forest were cut
down to make room for the Conestoga wagons and the gay stages
that swept through from Philadelphia to Pittsburg. These in their turn
became old-fashioned when the canal and Portage Railway were
done, and now we sit in a car that is like a palace, and think canals
and Conestogas very old stories indeed. In future generations swift
air ships may take the wonder away from the Empire State Express,
and make us listen unmoved when a man, standing in the station at
Philadelphia, calls the limited train for Pittsburg, Cincinnati, and St.
Louis.
 CHAPTER VIII
THE NATIONAL ROAD

The sea reaches inland almost to the northeast corner of the

state of Maryland. This long, wide arm of the ocean receives many
rivers and is known as Chesapeake bay. Near its north end is
Baltimore, one of the four great cities of our Atlantic coast. It is one
hundred and fifty miles from the open sea. If, instead of sailing up
the bay, we should turn toward the west, we could go up the
Potomac river, which is deep and wide. On our way we should pass
Washington’s estates at Mount Vernon, the old city of Alexandria,
and the national capital, Washington. We could not sail much farther
because there are falls in the Potomac which ships cannot pass. The
Potomac runs so close to Chesapeake bay that it is only forty miles
from Washington across to Baltimore.
Chesapeake bay is much like Delaware bay and the tidal
Hudson river, only it is larger than either. Baltimore is at a greater
distance from the open sea than Philadelphia is, and Philadelphia is
farther inland than New York, but each of these cities tried to get as
much of the western trade as it could.
The natural way for the men of Baltimore and Alexandria to go
across to the west was up the Potomac river and through its passes
in the mountains. But before they tried this they had settled much of
the low, flat land along the Potomac and about the Chesapeake in
Virginia and Maryland. This was often called “tide-water country,”
because the beds of the rivers are below sea level, and the streams
are deep enough for boats of some size.
 Fig. 33. Tollhouse West of Brownsville, Pennsylvania
When the land was first settled and the colonists found that they
could go almost everywhere by boat, they paid small heed to making
roads. They could visit their neighbors on other plantations and they
could load their tobacco and take it to market by the rivers. Many
plantations were beside rivers of such great depth that sailing
vessels bound for London could come up to the farmer’s wharf and
get his crop of tobacco.
In early days the members of the legislature were not always
given so much per mile to pay the stage fares between their homes
and the capital, but they were allowed the cost of hiring boats
instead. Many ferries were needed, and laws about them were made
before rules were laid down for bridges and roads. Several
lawmakers at one time would have been fined for their absence from
the legislature of the colony had they not been excused because
there was no ferry to carry them over the river which they would
have had to cross.
Around Annapolis “rolling roads” were made. These were wide
paths made as smooth as possible, in order that large hogsheads of
tobacco might be rolled, each by two men, to the market in that old
town.
 After a time the lowlands of the coast region began to fill up and
the people were pushing westward, just as they did in Pennsylvania
and New York. No man had so great a part in this westward
movement as the young surveyor, George Washington. In 1748 he
was sixteen years old, a tall, strong lad, full of courage and energy.
Lord William Fairfax, a rich English gentleman who had settled in
Virginia, had bought great tracts of forest land up the Potomac
behind the Blue Ridge mountains, and he was eager to have them
surveyed. Knowing that Washington had studied surveying, Fairfax
asked him to undertake the task. The boy consented; he went
beyond the Blue Ridge into the country along the Shenandoah,
camped in the woods, swam the rivers, toughened his muscles,
learned the ways of the red men, and three years later came back, a
grown man, ready for great things.
While Washington was getting his practice as a surveyor the
Ohio Company was formed to take up lands along the Ohio river,
and to keep the French from settling there. Lawrence, Washington’s
elder brother, was one of the chief men of this company. In 1753
Washington himself went west to the Ohio river. Day by day the
French were taking a firmer hold of that country, and Dinwiddie, the
old Scottish governor of Virginia, looked about for some one to carry
a warning letter to the commander of one of their new forts. The
messenger was also to keep his eyes open and report what the
French were doing on the upper waters of the Ohio. He chose
Washington, saying, “Faith, you’re a brave lad, and, if you play your
cards well, you shall have no cause to repent your bargain.”
Washington did not wait, but left on the day he received his
commission, late in October, 1753.
Christopher Gist, a famous frontiersman, was secured as guide,
and we can have no doubt that he and Washington formed a team,
ready to meet Frenchmen, red men, and the dangers of river and
forest. They made up their little party where the city of Cumberland,
Maryland, now stands. It is far up the Potomac, in the heart of the
mountains,—a long way beyond the Blue Ridge and the lands where
Washington had been surveying.
 At this place a large stream called Wills creek cuts through one
of the mountain ridges by a deep gorge and enters the Potomac. On
a hill, where these streams come together, was Fort Cumberland,
the great outpost of Virginia and Maryland. A fine church now stands
on the ground of the old fort, in the heart of the busy city of
Cumberland. This was the starting point for Washington’s expedition
and for many later ones into the western wilderness.
Washington made his dangerous journey with success. He
brought back a letter from the French commander, but of much
greater value was the story of all that he had seen. The colonists
now knew just what they would have to do to keep possession of the
Ohio lands.
It was not long before Washington went again as commanding
officer of a small army, and in 1755 he served under General
Braddock in the famous battle which resulted in the defeat of the
English and the death of their general. Washington, as we know,
brought off the troops with honor to himself. In each of these
expeditions something was done toward cutting away the trees and
grading a road from Fort Cumberland to the head of the Ohio river at
Pittsburg.
 Fig. 34. Milestone on the
Line of Braddock’s
Road, near Frostburg,
Maryland

On the line of Braddock’s road, a dozen miles west of

Cumberland, is a milestone, set up about a hundred and fifty years
ago. A photograph of it is shown above. It is a rough brown stone,
standing in a pasture half a mile outside the city of Frostburg, in
western Maryland. The stone was once taken away and broken, but
it has since been set up again and cemented into a base of concrete.
The view shows how it has been split up and down. On one side are
directions, and on the other are the words, “Our Country’s Rights We
Will Defend.”
Braddock’s journey from Alexandria to Fort Duquesne was an
uncomfortable one, to say nothing of its disastrous end. He bought a
carriage to ride in, but the road was not suited to a coach, as were
the roads he knew in old England. Beyond Cumberland, especially,
in spite of all the work his men could do upon it, it was so bad that he
was forced to take Washington’s advice and change the baggage
from wagons to pack horses.
 Gradually, as time went on, these rough paths were beaten down
into smoother thoroughfares. The same causes that led to the
development of the North were working also at the South. Along the
Potomac, as in New York and in Pennsylvania, the stream of colonial
life flowed westward. First the pioneers settled the lowlands around
Chesapeake bay and along the deep rivers; then as their strength
and courage reached beyond the mountains they found the forests
and fertile soil behind the Blue Ridge. Farther within the rugged
highlands they built Fort Cumberland and sent out discoverers and
armies to the Ohio river. When the woods were cleared and towns
and states grew up on the Ohio, there was frequent occasion to
cross the mountains for trade, for travel, and to reach the seat of
government, which in 1801 was moved to Washington on the
Potomac.

Fig. 35. Old Road House, Brownsville, Pennsylvania

These glimpses of colonial journeys will help us to understand
why the National Road came to be built. About one hundred years
ago the government began to take a great interest in opening roads,
especially across the Appalachian mountains, to Ohio, Kentucky,
and other parts of the Mississippi valley. Washington, who died in
1799, had said much about this work, for he not only wanted western
trade to come to Virginia instead of going to New Orleans, but he
also felt that so long as the mountains kept the East and the West
apart we should never have one common country, held together by
friendly feelings.
The people of Baltimore, like those of New York and
Philadelphia, were eager to have the best road to the West, that their
business might be benefited. Not far from Baltimore is an old place
called Joppa, and several roads are still known as “Joppa roads.”
The town is older than Baltimore and was once the chief trading
town in the northern part of Maryland; but Baltimore was well
situated on an arm of the great bay, and by this time had gone far
ahead of its old rival.
A number of good roads had been built in Maryland, among
them a famous one leading out westward to Frederick. This was in
the direction of Hagerstown, and still farther west was Cumberland.
The United States government decided to build a great road to Ohio,
beginning at Cumberland. To get the benefit of this, the men of
Baltimore went to work to push the Frederick pike westward to the
beginning of the National Road.
So it came about in 1811 that the first contracts were let for
building parts of the National Road. We remember that the Erie
canal was not started until six years later. The act of Congress which
ordered the making of the road provided that a strip four rods wide
should be cleared of trees, that it should be built up in the middle
with broken stone, gravel, or other material good for roads, and that
all steep slopes should be avoided. The road was opened to the
public in 1818, one year after the Erie canal was begun. The original
plan was to make it seven hundred miles long, reaching from
Cumberland to the Mississippi river, but it was never carried out.
The Maryland roads, as we have seen, ran west from Baltimore
and Washington to Frederick, east of the Blue Ridge; to Hagerstown,
in the Great Valley; and to Cumberland, in the mountains.
Cumberland is a stirring town of about twenty thousand people, and
with its great business in coal, iron, and railroads it seems like a
larger city. Thence the National Road runs through the gap in Wills
mountain (Fig. 36) to Frostburg, a dozen miles west and fifteen
hundred feet higher. The road soon bears northward into
Pennsylvania and crosses the Monongahela river at Brownsville,
about forty miles south of Pittsburg. Coal is mined here, and boats
were running in those early days, as coal barges and steamboats
run to-day, down to the great iron city.
From Brownsville the pike leads over the hills and comes down
to the Ohio river at Wheeling, West Virginia. It then passes on
through Ohio, touching Columbus, the capital, on the way to Indiana
and the Mississippi.
We sometimes admire the cars marked with the sign of the
United States post office, which we see drawn by a swift locomotive
at a speed of sixty miles an hour; but when the government put its
mail coaches on the National Road from Washington to Wheeling, no
doubt they seemed quite as wonderful to the people of that time. And
it was only twenty-five years since the people of Utica had thought it
so remarkable that six letters had come to them in one mail! Soon
passenger coaches were rushing along at ten miles an hour, and
sometimes even faster. There were canvas-covered freight wagons,
each of which carried ten tons, had rear wheels ten feet high, and
was drawn by twelve horses. In those days life was full of stirring
interest on the National Road.
 Fig. 36. Cumberland and the Gap in Wills Mountain
There were rates of toll for all sorts of animals and wagons. The
toll was higher for hogs than for sheep, and more was charged for
cattle than for hogs. If the wagons had very wide tires, no toll was
demanded. Drivers sometimes lied about the number of people in
their stages, so as to pay less toll. The stages were not owned by
the drivers but by companies, which bid for travelers and freight, as
railways do now. There were penalties for injuring milestones or
defacing bridges, showing that some people then were like some
people now. The companies had interesting names. There were the
“Good Intent,” “Ohio National Stage Lines,” the “Pilot,” “Pioneer,”
“June Bug,” and “Defiance.” Not one of them cared for mud or dust,
for horses or men, if only it could be the first to reach its destination.
There must have been dust enough, for twenty coaches with their
many horses sometimes followed one another in a close line.
 Fig. 37. Bridge and Monument, National Road, near
Wheeling, West Virginia
Henry Clay was one of the chief advocates of this road, and a
monument built in his honor may be seen near the bridge, shown in
Fig. 37. It is a few miles east of Wheeling. At Brownsville a small
stream called Dunlap’s creek flows into the Monongahela from the
east. Over it is an iron bridge on the line of the National Road.
According to a story told in Brownsville, Henry Clay was once
overturned as he was riding through the creek before the bridge was
built. As he gathered himself up he was heard to say, “Clay and mud
shall not be mixed here again.” The story goes that he went on
immediately to Washington and got an order for the building of the
bridge.
Whether this be true or not, it is certain that he and many other
statesmen traveled over the National Road. They could not have
private cars, nor did they go in drawing-room coaches, as we can if
we choose. Anybody might chance to sit beside these men of
national fame, as day after day they rode through the valleys and
over the mountains, stopping at the wayside hotels for food and rest.
Some of the old hotels, tollhouses, and bridges, as they look to-
day, are shown in the illustrations in this chapter. The road itself was
long ago given up to the different states and counties through which
it runs, but it still tells to the traveler who goes over it many a story of
the life of a hundred years ago.
 CHAPTER IX
THE BALTIMORE AND OHIO RAILROAD

Even after the Erie canal was built and long lines of boats were
carrying the grain and other products of the West to New York, the
men of Virginia and Maryland did not give up the notion of still
making the trade of the western country come their way. They
planned the Chesapeake and Ohio canal, to reach the Ohio river,
and thought that other canals across the state of Ohio would let them
into lake Erie. By the Ohio river they would connect with New
Orleans and the upper Mississippi river, and through lake Erie they
could reach the towns and farms that border lake Huron, lake
Michigan, and lake Superior.
A canal along the Potomac valley had been talked of several
years before the Revolution, when Richard Henry Lee laid a plan for
it before the Assembly of Virginia. Doubtless others thought of it too,
as of the Erie canal, long before it was made. At the end of the War
of the Revolution Washington made a long journey into the wild
woods of New York. He went to the source of the Susquehanna at
Otsego lake, visited the portage between the Mohawk and Wood
creek, and saw for himself that New York had a great chance for
navigation and trade. But he had a natural love for his own Virginia,
and he did not intend to let New York go ahead of his native state.
His journeys across the mountains as a surveyor and as a soldier
gave him a knowledge of the Ohio country, and as he had himself
taken up much good land there, he wished to have an easy way, by
land or water, from the sea to the rich Ohio valley. So he thought
much about a canal to run by the side of the Potomac, and he joined
with others who felt as he did to form the Potomac Company. They