OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

C H A PT ER 1

A gentle introduction to genomics

 Key points
• The “central dogma” of biology states that genomic information flows from DNA to RNA to protein.
• Most human DNA does not encode genes (i.e. is non-coding) but includes various regulatory elements that facilitate the
regulation of gene activity.
• Mutations can have functional consequences for both coding regions (e.g. changes to protein sequence) and non-coding
processes (e.g. altering the regulation of the protein).
• The most common form of genetic variation in humans is single nucleotide polymorphisms (SNPs), and inheritance of
SNPs is not typically random and often correlated.

1.1 Introduction that defines your unique biology both as an indi-

The motivation for writing this book stems from a
strong belief that anyone who has an interest in
exploring their personal genome has a right to gain
access to the tools and knowledge required to do so.
Therefore, we expect (and hope!) that some readers
of this book may lack a formal education in genetics
or genomics. Although we have made a strong
attempt to provide appropriate explanations and
background knowledge to enable a thorough under-
standing of the materials in each individual chapter,
an understanding of fundamental concepts in
genomics will enable the reader to utilize the knowl-
edge and tools represented in this book optimally.
In this chapter, we provide a brief and gentle intro-
duction to fundamental concepts in genomics aimed
at providing a fundamental understanding of the
human genome, which will greatly enhance one’s
ability to analyze and interpret a personal genome.
1.2 What is a genome?
A genome is quite simply the entire sum of the
genetic components of a living species. Plainly, you
can think of your genome as all the “genetic stuff”
vidual within the human species, as well as the
member of a distinct species within the animal king-
dom. Popular media often refers to the genome as
the set of genes unique to an individual, but we will
see that a genome is much more than genes. In fact,
only ~2% of the human genome is known to code
for actual genes; until recently, the remaining 98%
was referred to as “junk” DNA, however this
moniker is quickly fading as we learn more and
more about the important functional roles of this
supposed “junk” DNA. This chapter provides an
introductory minimum of knowledge about the
human genome and the associated molecular biol-
ogy required for informed analysis and interpreta-
tion of a personal genome.
1.2.1 The architecture of a human genome
The human genome is fundamentally composed of
deoxyribonucleic acid (DNA), which is composed
of a set of four distinct nucleotides (“bases”) that
combine according to specific pairing rules along a
sugar-phosphate backbone, which gives the genome
its well-known double-helix conformation. The
four bases in the human genome are adenine (A),

Exploring Personal Genomics. First Edition. Joel T. Dudley and Konrad J. Karczewski.
© Joel T. Dudley and Konrad J. Karczewski 2013. Published 2013 by Oxford University Press.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

4 EXPLORING PERSONAL GENOMICS

Phosphate Molecule gle (unpaired) strand of nucleotide bases (e.g., ATG-

TAACGT) since the complementary strand can
Deoxyribose
Sugar Molecule always be inferred from the purine-pyrimidine
pairing rule. Given that the entire genome can be
encoded using only four characters representing the
Nitrogenous
Bases
four nucleotides, you may be wondering how we
can identify the protein-coding DNA regions (i.e.
genes) from the non-coding DNA. As we will learn
in the following sections, additional genomic fea-
T
A tures, such as the higher-order patterning of nucle-
otide bases in the genome, and the organization of
DNA in the cell nucleus, encode and govern the
complex functionality of the human genome.
C
G

1.2.2 Structure of protein-coding regions

C of our DNA resides in genes, which are broadly
G
T is transcribed into messenger RNA (mRNA) and
A
Weak Bonds
Between
Bases
Sugar-Phosphate
Backbone
Figure 1.1 Molecular structure of DNA: DNA is made up of four base
pairs (A, T, C, and G) attached to a sugar-phosphate backbone. The two
strands of DNA are formed by weak hydrogen bonds, where A pairs
with T and C pairs with G. DNA is “complementary”, meaning one
strand can be reconstructed from the other by reversing it and switching
A with T and C with G. Image courtesy of the U.S. Department of Energy
Genomic Science program (<http://genomicscience.energy.gov>). Image
reprinted from Molecular Evolution and Phylogenetics by Nei and
Kumar, OUP: 2000.
thymine (T), guanine (G), and cytosine (C). To form
the double-helix structure, two DNA strands “pair”
(through weak hydrogen bonding), and due to
physiochemical constraints, A always pairs with
T, and C always pairs with G (Figure 1.1). Because
of this strict pattern of complementarity, it is usually
sufficient to represent the human genome as a sin-
As previously mentioned, only a small fraction
defined as the functional units of the genome. These
regions typically “code” for proteins (with excep-
tions discussed below), wherein the DNA sequence
translated into proteins (Figure 1.2). The intermedi-
ate step, RNA (ribonucleic acid), is a molecule that
is quite similar to DNA in structure, except with the
addition of a hydroxyl group (-OH), and thymine
(T) is replaced by a similar nucleic acid, uracil (U).
Additionally, RNA is most often single-stranded.
The final products are the proteins, which make up
the various constituent parts of cells in the human
body, including structural proteins (that hold the
cell together), enzymatic proteins (that catalyze cel-
lular reactions), and signaling proteins (that com-
municate signals within or across cells). It is these
protein-coding genes that are the most studied
regions in the human genome (and have been most
prominently described in the popular media), and
thus, they form the basis of our classical under-
standing of human genetics and genomics.
Unlike the protein-coding regions of many micro-
organisms, such as bacteria, the protein-coding
regions of human DNA are not contiguous, but
instead are interrupted with non-coding DNA. The
protein-coding regions are fragmented, where the
information that encodes a gene is broken up into
chunks called exons, which are interspersed with
non-coding nucleotide regions called introns (see
Figure 1.2). When genes are expressed in human

DNA
GT AG GT AG
5’-regulatory
region Transcription
Exon 1 Exon 2
Pre-mRNA
Intron 1 Intron 2
5’-untranslated Introns 3’-untranslated
region spliced out
Exon 1 Exon 2 Exon 3
Processed mRNA
Translation
Polypeptide
Figure 1.2 The central dogma of biology: DNA is “transcribed” into
RNA, which is then processed and “translated” into protein. Gene
regions are transcribed into a pre-mRNA (pre-messenger RNA), which
includes untranslated regions (at the 5’ and 3’ ends) as well as
introns. The introns are “spliced out” in a processing step, before the
processed mRNA is translated into a protein. Image reprinted from
Molecular Evolution and Phylogenetics by Nei and Kumar, OUP: 2000.
cells, the cellular machinery removes (“splices out”)
the intronic regions and rejoins the remainder to
form messenger RNA. We will learn later how the
fragmented structure of protein-coding genes ena-
bles additional functional capacity through a mech-
anism known as alternative splicing, where exons
are occasionally skipped in different combinations
to create alternative gene products called protein
isoforms. Finally, these exons and introns are flanked
by “control regions” for the gene, known as pro-
moters. These regions are involved in regulating
gene expression, as they encode information about
when and how much of the gene should be
expressed or transcribed into mRNA. For exam-
ple, these regions may control whether a gene
should be expressed in liver or brain or in response
to a stimulus, such as a disease.
1.2.3 The myth of junk DNA
Even before the human genome was initially
sequenced, the discovery that most (over 95%) of
the human genome did not directly code for pro-
teins led to the common usage of the phrase “junk
DNA” to refer to this large proportion of human
DNA with unknown function. Recent advances in
genome science have proven this phrase to be a
misnomer, with scientists regularly decoding the
OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 5
nature and function of various aspects of this large,
3’-end non-coding expanse in the human genome using
new tools and technologies. By comparing the
genomes of many species, we have hints from evo-
Exon 3
lution that some of this DNA may be functionally
important, as 4–7% of the human genome is present
region
across mammals, a concept known as evolutionary
conservation. It is likely that advances in func-
tional genomics in the near future will offer insights
into the precise nature and function of elements in
these large non-coding regions. A breakdown of
the DNA contained in the human genome is shown
in Figure 1.3, but some of the key genomic ele-
ments known to comprise these non-coding regions
include:
• Gene-related sequences: These include introns,
untranslated regions (UTRs), and promoter
regions that may have numerous roles in the regu-
lation of gene expression. Introns can be “spliced
out” to generate new gene products from the same
exons (only in different combinations; see Box 1.2
“Alternative Splicing”). UTRs and promoters are
involved in regulating gene expression levels by
affecting transcription and degradation of mRNA
(see Section 1.6).
• Non-coding RNAs (ncRNAs): ncRNAs are RNA
molecules that are transcribed, but not translated,
with the actual RNA molecule itself performing
a functional role. These appear to have diverse
roles across the genome, such as regulating the
expression of genes (e.g. miRNAs, see Section 1.4)
and serving roles in the structural integrity of the
chromosome (e.g. telomerase RNA).
• Psuedogenes: Pseudogenes are remnants of
ancient genes that were once functional in an
ancestral species, but have now lost their func-
tion (i.e. they are switched off) and are accumu-
lating genetic “noise” in the form of random
mutations.
• DNA repeats: Repetitive patterns of DNA
nucleotides of various length sometimes referred
to as satellite DNA. These repeats are thought to
have come about through “slippage” errors in
DNA replication, however some have speculated
that these repeat elements have important roles in
shaping species evolution.
• Retrotransposons: Recent evidence has dem-
onstrated that up to 8% of the human genome
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

6 EXPLORING PERSONAL GENOMICS

Human genome
3200 Mb

Genes and gene-related Intergenic DNA

sequences 1200 Mb 2000 Mb

Other
Interspersed
Related intergenic
Genes repeats
sequences regions
48 Mb 1400 Mb
1152 Mb 600 Mb

Microsatellites
Gene Introns. 90 Mb
Pseudogenes
fragments UTRs
Various
LINEs LTR elements 510 Mb
640 Mb 250 Mb

SINEs DNA transposons

420 Mb 90 Mb

Figure 1.3 Breakdown of the human genome: While a small fraction of the human genome is directly translated into protein, the overwhelming
majority does not. Some of this “non-coding” DNA, especially DNA near genes, is involved in the regulation of gene expression. Other segments
are repetitive elements, including long- and short-interspersed DNA elements (LINEs and SINEs), long terminal repeats (LTRs), and transposons.
Image reprinted with permission from Genomes 2 by Brown.

has been formed by retrotransposon DNA from
Human Endogenous Retroviruses (HERVs),
although the human genome contains a larger
proportion (~25%) of DNA that is identifiable as
being formed by retrotransposons. These viral
retrotransposons are suspected to have had a
significant effect on the evolution of the mamma-
lian genome, and particularly in the evolution of
humans and other hominids.
1.2.4 Packaging Genomic DNA into
Chromosomes
If you were to stretch a human genome out into a
single contiguous strand, you would find that it
would be nearly 1.83 meters (6 ft) from end to end.
Every cell in the human body with a nucleus con-
tains a full copy of an individual’s genome, which is
packaged inside the cell in the form of densely
packed structures called chromosomes (see
Figure 1.4). The number of chromosomes used to
package the genome into a cell nucleus varies by
species and human cells have 23 chromosome pairs
(46 total chromosomes), with one of these pairs
being a sex-determining chromosome pair (XX in
females and XY in males). One member of each of
these chromosome pairs is inherited from the
mother, and the other is inherited from the father. In
addition, the mitochondrial genome is a small circu-
lar genome that is inherited only from the mother
(see Box 1.1).
As such, each cell in our bodies is a diploid cell,
with two copies of each chromosome. Gametes, or
reproductive cells (i.e. sperm and egg), are the only
exception to this rule. Human gametes are haploid
cells, which contain half of the typical amount, or 23
chromosomes. This is necessary because gametes
from two individuals will fuse in the process of sex-
ual reproduction to form a zygote, which will
develop into an embryo with 46 chromosomes per
cell. Therefore, chromosomes are not only impor-
tant for the packaging of DNA into the cellular
nucleus, but also as structures and units of genetic
inheritance in human sexual reproduction.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 7

Chromosome
Nucleus Chromatid Chromatid
Telomere

Centromere

Telomere
Cell

Base Pairs Histones

A
T G
C A DNA
T C
T (double helix)
G
G G
G
A
C

Figure 1.4 DNA packaged in chromatin: DNA is wrapped around histones, which is organized into chromatin. This chromatin makes up
chromosomes, which are found in the nucleus of the cell. The cell efficiently packs the 1.8 meters of linear DNA into a nucleus that is 6 microm-
eters in diameter. Image courtesy of the National Human Genome Research Institute (artist: Darryl Leja).

Chromosomes are composed of a complex bun-
dling of genomic DNA and proteins known as chro-
matin. The major protein constituents of chromatin
are a special set of proteins called histone proteins,
around which genomic DNA is wrapped. It can be
helpful to visualize the relationship between DNA
and histone proteins using a thread and spool anal-
ogy. In this way, you can think of DNA as a thread
of string that is wound tightly around a histone
“spool” (146 base pairs at a time to be exact!). When
the DNA is wrapped around this histone “spool”,
the resulting particle is called a nucleosome. This
process is repeated along the length of genomic
DNA for a particular chromosome, creating a “beads
on a string” structure with a contiguous strand of
nucleosome “beads” on a genomic “string.” These
nucleosome “beads” are connected by another type
of histone (known as linker histones) that organizes
the nucleosomes into a more tightly-packed struc-
ture known as chromatin fiber. These chromatin fib-
ers form the basis of chromatin, which can be further
classified as heterochromatin, a condensed form, or
euchromatin, an extended form. Thus, human
genomic DNA is packaged into cellular nuclei in a
condensed and complex structure, as opposed to
floating around as one long strand. The packaging
of genomic DNA in the chromatin structures of the
nuclei can have a significant effect on the expression
of genes (See Section 1.4).
In human cells, during mitosis, a chromosome
duplicates and condenses such that it can be
visualized under a microscope (Figure 1.5). The
duplicated pair (also known as “sister chromatids”)
is physically joined by a structure known as a
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
8 EXPLORING PERSONAL GENOMICS
Telomere
Short arm (p)
Centromere
Long arm (q)
Telomere
Figure 1.5 Schematic of a chromosome: The two sister chromatids
(perfectly identical copies) of a chromosome can be visualized after it
is duplicated during mitosis. These copies are connected off-center by
a centromere, resulting in the long and short arms (p and q arms),
which contain telomeres at their ends.
centromere, giving chromosomes their characteris-
tic crisscross appearance (although in reality the
chromosomes are not crossed over each other when
joined by the centromere, but rather “pinched”
together). Chromosome pairs are joined by the cen-
tromere in an asymmetrical fashion, such that the
unconnected region or arm on one sides of the cen-
tromere is typically shorter than the other. The short
arm of a joined chromosome pair is known as the
“p” arm and the longer arm is known as the “q”
arm. This convention is often used in annotating
specific coordinates of a genomic region (e.g. 6q ori-
ents to the long arm of chromosome 6). The regions
at the end of each chromosome are called telomeres.
Telomeres are composed of repetitive sequences of
non-coding DNA that protects the chromosome
from damage during cell division. Each time a cell
divides, the telomeres are shortened, and this short-
ening has been implicated in some age-related
diseases.
1.3 How does a genome work?
Having a particular gene encoded in your genome
is largely inconsequential unless it is “expressed” to
result in some biological effect inside or outside the
cell. Specifically, while your genes are encoded by
your genomic DNA, they generally do not exert any
influence until they are expressed through a process
called transcription, which eventually leads to the
formation of functional protein products through a
Box 1.1 “The Mitochondrial Genome”
Many people will be surprised to hear that the genome
found in the nuclei of our cells is not the only genome
found in the human body. Mitochondria, which are
energy-producing sub-cellular components found in most
cells in the human body, have their own genome. In addi-
tion to being much smaller than the human genome, con-
taining just 37 genes across ~16,569 nucleotide base
pairs, mitochondrial DNA looks quite a bit different than
human genomic DNA. Instead of being packaged into
chromatin by histones, mitochondrial DNA takes a circular
form and it does not contain any introns. These character-
istics appear to make the mitochondrial genome more
similar to bacterial genomes than the human genome. In
fact, the leading theory on the origins of the mitochon-
drial genome is the endosymbiont theory, which asserts
that mitochondria developed from bacteria that were
taken up by an ancient eukaryotic cell and subsequently
developed a symbiotic relationship with the host cell.
There is a substantial amount of evidence supporting this
theory. For example, the ribosomal RNA (rRNA) found in
mitochondria is more similar to the rRNA found in bacte-
ria rather than mammals.
From the perspective of personal genomics, there are
two key properties of mitochondrial DNA that useful to
understand. First, since the mitochondria are the “power
plants” of the cell, mutations in mitochondrial DNA can
often lead to a host of genetic disease conditions in
humans. Second, mitochondrial DNA is only inherited
from the mother, and is therefore useful for investigation
of inheritance and ancestry along an individual’s maternal
lineage. If you push this idea of maternal inheritance to its
extreme conclusion, you can assert that the mitochondrial
DNA found in the present human population could be
traced back to a single female individual, who is popularly
referred to as “Mitochondrial Eve.” Molecular evolution-
ary studies of human mitochondrial DNA estimate that
Mitochondrial Eve most likely lived around 200,000 years
ago and most likely in East Africa. In Chapter 5, we will
learn how to map mitochondrial DNA haplogroups back
to particular human lineages.
process called translation. While the process of
expressing a single gene to make a single protein
product is a tremendously complex process in its
own right, the cell must be able to precisely express
thousands of genes in concert to support that viabil-
ity of a just single cell, much less a complex, multi-
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
cellular species such as humans. To complicate
matters further, cells must also achieve a specific
timing of this regulation: for instance, fetal haemo-
globin is only expressed in early development and
then switched off to make way for adult haemo-
globin. This symphony of gene expression is
achieved through a process known generally as
gene regulation.
1.3.1 From genes to proteins: the central
dogma of molecular biology
This sequential transfer of information from the
genome to proteins is called the central dogma of
molecular biology, a phrase that was coined by
co-discoverer of DNA Francis Crick in 1958. The
central dogma of molecular biology states that
flow of genetic information in molecular biology is
DNA → RNA → Protein, or simply that DNA is
transcribed to RNA and RNA is translated into
proteins (Figure 1.2). Like many rules of thumb,
this is an oversimplification the relationships
between DNA, RNA and proteins, but it is a useful
tool for understanding the general relationship
between these molecules and the flow of informa-
tion that starts at the genome and produces protein
products with biological activity. Despite their fun-
damental nature, aspects of gene transcription and
translation are still areas of active research; for
more information, we refer readers to “Recom-
binant DNA: Genes and Genomes” by Watson,
Caudy, Myers, and Witkowski and other reviews
such as [Chen and Rajewsky, 2007].
1.3.2 Gene transcription
One of the first major steps in gene expression is tran-
scription, which creates an RNA copy of a gene, called
messenger RNA (mRNA), from its genomic DNA
template (Figure 1.6). The actual synthesis of mRNA
is performed by an enzyme called RNA Polymerase
(RNAP). In order to transcribe a gene into an mRNA
transcript, RNAP must bind to genomic DNA at the
beginning of a gene sequence, known as a promoter
region, which lies in the “upstream” or 5’ (pronounced
“five-prime”) position.
Since RNAP does not know the location of par-
ticular genes in the genome or when the cell needs
A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 9
to express a particular gene, it is aided by proteins
called transcription factors, which recognize short
sequence motifs in gene promoter regions. These
transcription factors then recruit RNAP to tran-
scribe an RNA copy of the gene. During this proc-
ess, the RNAP will move along the double-stranded
DNA in the 5’ → 3’ direction of the coding strand
and it will use the complementary strand as a tem-
plate (read in the 3’ → 5’ orientation) to encode an
mRNA transcript of a gene (Figure 1.6). Recall that
DNA base pair complementarily ensures that a
transcript encoded from the complementary
strand will match the information found in the
“intended” coding strand. Note that the nucle-
otide code used by mRNA is slightly different
than that of DNA: any positions encoded with a
thymine (T) in DNA will be encoded using a uracil
(U) in RNA.
However, the mRNA strand produced by the
RNAP does yet not represent a viable representa-
tion of a gene, as it will contain non-coding intronic
regions in addition to the desired exonic regions.
This immature mRNA strand is known as precur-
sor mRNA (pre-mRNA), which must have the non-
coding intronic regions removed in a process called
splicing, before it can be used to create a viable
protein product. This arrangement of exons can
Genomic DNA template
Transcription Nucleus
RNA processing
RNA transcript
Cytoplasm
mRNA
Protein Translation
tRNA
polypeptide
Ribosome
Figure 1.6 Transcription and translation: DNA is transcribed into
RNA transcripts (pre-mRNA) which is processed to mature mRNAs in
the nucleus. These mRNAs move from the nucleus to the cytoplasm,
where the translational machinery (the ribosome along with tRNAs)
translates the message into a protein.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

10 EXPLORING PERSONAL GENOMICS

data in the 5´ to 3´ direction. The information regard-

Box 1.2 “Alternative splicing”
The alternative combination of exons during pre-mRNA
processing is known as alternative splicing (see
Figure 1.7). As you might imagine, skipping certain exonic
gene regions can have the consequence of the same ini-
tial gene sequence producing very different protein
sequences. Note that in alternative splicing, the order of
exons is still maintained, but which ones are used to pro-
duce the final mRNA molecule may vary. Alternative splic-
ing is facilitated by a number of complex genetic and
biochemical factors, driven by a splicing machinery that
can recognize and bind to particular regions in the gene.
Additionally, there are a number of other factors affecting
the resulting alternatively spliced transcript, such as the
retention of particular introns, or a mutually exclusive
relationship between certain exons. For the purposes of
this book, the main takeaway point from alternative splic-
ing is that a single gene sequence encoded in the genomic
DNA can in fact produce a repertoire of differing protein
products.
produce a large amount of genetic diversity, if cer-
tain exons are skipped, forming various combina-
tions of exons in the mature mRNA (see Box 1.2).
ing the precise composition of the protein encoded by
the mRNA is represented in a genetic language that
uses a triplet of nucleotide sequences, or a codon, to
encode one of the 20 amino acids used by mammalian
cells to build proteins. Each codon is recognized by a
tRNA (transfer RNA) molecule that carries with it an
amino acid, which is added to a peptide chain by the
ribosome. The codon translation used in human cells
is shown in Table 1.1.
Many of the amino acids are encoded by multiple
codons; when a set of codons encodes the same
amino acid, these codons are said to be synony-
mous, and provide redundancy in the genetic code.
There are several theories as to why such redun-
dancy exists in the genetic code, but leading theo-
ries suggest that this redundancy protects from the
effects of mutation. There are also a few special
codons that serve as regulatory signals to the trans-
lational machinery. The start codon (AUG in RNA)
encodes the amino acid methionine and serves as a
signal that “tells” the translational machinery to
begin protein synthesis. The stop codons (TGA,
TAG, and TAA) serve as a signal that then end of a
protein has been reached and that protein synthesis
by the ribosome should stop.

1.3.3 Gene translation

Table 1.1 Codon table: With only a few exceptions, the genetic code
After the pre-mRNA is processed by the spliceosome, of life is shared among organisms. Each set of three nucleotides
and a mature mRNA is produced, the cell transports (which form a codon) code for a specific amino acid. Note that there
the mRNA out of the nucleus into the cell’s cytoplasm, is a large degree of redundancy in this code (e.g. CCU, CCC, CCA, and
CCG all code for proline), the result of which is that certain mutations
where the information encoded by the mRNA can occur without affecting protein structure
sequence to synthesize a protein (Figure 1.6). The syn-
Second letter
thesis of proteins from mRNA by the translational U C A G
machinery in the cell is perhaps one of the most amaz- U
UUU Phe UCU UAU Tyr UGU Cys
ing examples of molecular machinery yet revealed by U UUC UCC Ser UAC UGC C
UUA UCA UAA Stop UGA Stop A
science. The same molecular process that translates UUG Leu UCG UAG Stop UGG Trp G
information from mRNA into proteins is capable of
CUU CCU CAU His CGU U
producing the soft proteins that make up human hair, C
C CUC Leu CCC Pro CAC CGC Arg
A
as well as the tough collagen fibers and matrix pro- CUA CCA CAA CGA
Third letter
First letter

CUG CCG CAG Gin CGG G

teins that make up the hard human endoskeleton. The
AUU AUU AAU Asn AGU Ser U
translation of mRNA information into proteins is C
A AUC Ile AUC Thr AAC AGC
A
facilitated by a complex macromolecule called the AUA AUA AAA AGA
AUG Met AUG AAG Lys AGG Arg G
ribosome, a complex made up of various protein and
RNA subunits. The ribosome binds to the mRNA in GUU GCU GAU Asp GGU U
GUC GCC Ala GAC GGC Gly C
the cellular cytoplasm at the 5´ (upstream) region of G GUA Val GCA GAA GGA A
GUG GCG GAG Glu GGG G
the mRNA sequence and proceeds to read the mRNA
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 11

Exon 1 Exon 2 Exon 3 Exon 4 Exon 5

DNA
Exon 1 Exon 2 Exon 3 Exon 4 Exon 5
RNA

Alternative Splicing

1 2 3 4 5 1 2 4 5 1 2 3 5
mRNA
Translation Translation Translation

Protein A Protein B Protein C

Figure 1.7 Alternative splicing: After DNA is transcribed into an RNA transcript, the relevant exons are chosen to form mature mRNAs. The choice
of which exons to use can result in different mRNAs and different protein products. Note that the order of the exons is kept intact. Image courtesy
of the National Human Genome Research Institute.

1.3.4 Protein folding and 3-D structure such as G-protein coupled receptors (GPCR), which
serve as targets for the majority of drugs on the mar-
When a protein polymer is being synthesized by the
ket. Protein structures are usually solved by scientists
ribosome during translation, it does not persist as a
in the field of structural biology, where they employ a
long string of amino acids chained together. Instead,
technique called X-ray crystallography; however, this
the protein sequence immediately begins to fold into
approach is complex and time-consuming. In the effort
complex three-dimensional (3-D) structures (Figure
to accelerate the elucidation of human protein 3-D
1.8). The rules governing how a protein will fold are
structures, many hope to develop computational
profoundly complex and incorporate a large number
approaches to predicting 3-D protein structure from
of parameters. Certainly, the amino acid sequence itself
the 2-D amino acid sequences of human proteins,
will affect protein folding, where the polarity, hydro-
which are known. If you are interested in assisting the
phobicity, covalent bonding, and many other physio-
effort to identify the 3-D structures of human proteins,
chemical parameters of the amino acids will influence
you might consider joining the Folding@Home effort
the folding dynamics. Protein folding is also influenced
from Stanford University (<http://folding.stanford.
by many factors extrinsic to the amino acid sequence,
edu>), which allows you to download protein-folding
such as the polarity of the solvent in which it is folding,
software that runs on your computer and allows you
solvent ion concentration, temperature, and the pres-
to contribute your unused computing power to
ence of special protein folding “helper” molecules
advanced protein-folding algorithms.
called chaperones. Ultimately, the protein will fold into
a 3-D structure that minimizes the free energy across
the molecule (i.e. is chemically and energetically sta- 1.4 Gene regulation: when and where a
ble). As you might imagine, errors in protein folding
gene is expressed
can cause severe disruptions to the proteins function,
and indeed protein misfolding lies at the heart of many Gene regulation encompasses many of the underly-
well-known diseases, such as Alzheimer’s disease. ing processes discussed so far, such as transcription
The true 3-D structures of many human proteins are and translation (Figure 1.9). However, in this sec-
yet to be determined, even for very important proteins tion, we describe some of the higher-level processes
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
12 EXPLORING PERSONAL GENOMICS
Levels of protein organization
Primary protein structure
is sequence of a chain of
Amino Acids amino acids
Pleated sheet Alpha helix
Secondary protein structure
occurs when the sequence of
amino acids is linked by
hydrogen bonds
Pleated
sheet
Tertiary protein structure
occurs when certain attractions
are present between alpha
Alpha helices and pleated sheets.
helix
Quaternary protein structure
is a protein consisting of more
than one amino acid chain.
Figure 1.8 Protein folding: Once an mRNA is translated into an
amino acid peptide, the primary structure (sequence) must be folded
into a functional protein. Secondary structure, including alpha helices
and beta sheets, are formed and folded across each other to form
tertiary protein structures. Additionally, multiple subunits of a protein
may come together to form a quaternary structure. Image courtesy of
the National Human Genome Research Institute.
that govern the initiation and termination of these
processes as part of more complex “programs” gov-
erning the precise expression of a gene or genes. The
biology of gene regulation is still an area of intensive
research, and a comprehensive look at the complex
phenomena underlying gene regulation has filled
the contents of entire books. Here, we briefly intro-
duce several key concepts in gene regulation that
provide the necessary biological background for
understanding concepts that are discussed later in
this book. Interested readers are strongly encour-
aged to read more on this topic, as knowledge of this
subject is critical to the study of the genetic basis of
human disease. One excellent resource for further
study is the freely-available edition of Genomes 2
which can be found online at the NCBI Bookshelf.
Post-transcriptional control
MicroRNAs, alternative splicing,
alternative polyadenylation,
RNA-binding proteins, etc.
miRNAs
TF TF mRNA
Transcriptional control
Transcription factors, chromatin state,
combinatorial control, co-factors,
alternative promoters, etc.
Figure 1.9 Overview of gene regulation processes: While the coding
segments of genes take care of the structure of the protein, other
processes regulate the “expression” of the gene, or the amount
present in a cell at a given time. Regulatory mechanisms at the
transcriptional level include binding of transcription factors to
promoter regions, as well as chromatin remodeling and state (histone
modifications). Additionally, the gene can be regulated at the
post-transcriptional level through alternative splicing, microRNA-
mediated regulation, and other chemical modifications of the mRNA.
Reprinted by permission from Macmillan Publishers Ltd: Chen, K. &
Rajewsky, N. The evolution of gene regulation by transcription factors
and microRNAs. Nature Review Genetics 8, 93–103 (2007).
• Transcription factors: A set of regulatory proteins
that contain special functional features in their
proteins structures called DNA binding domains,
such as zinc finger and leucine zipper domains.
These DNA binding domains allow the proteins
to bind to specific regions of the genome, called
recognition sequences, which are typically adja-
cent to the genes that they regulate. Once a tran-
scription factor binds to a recognition sequence
upstream of a gene, a myriad number of regula-
tory events can happen that might affect the gene’s
expression. The transcription factor could initiate
transcription and therefore instigate expression
of the gene, or it may repress the gene by block-
ing the transcriptional machinery. DNA-bound
transcription factors will typically interact with a
vast array of additional modifier proteins, such
as coactivators, corepressors, or kinases which
may alter the activity of the transcription factor,
or form larger regulatory complexes.
• Promoters: Regions of regulatory DNA to which
transcription factors typically bind. Promot-
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
ers usually lie upstream of genes and are often
composed of specific patterns of DNA base pairs
that are recognized by transcription factors. For
example, the sequence TATAAA is known as a
TATA box motif which is recognized by TATA
binding protein transcription factors. When tran-
scription factors are bound to promoters, they
will typically form complexes that attract RNA
Polymerase II to bind to the promoter region and
initiate transcription of the adjacent gene.
• Chromatin remodeling: Recall that the DNA in
human cells is packaged into a highly condensed
structure called chromatin. The chromatin state
of a particular genomic region can regulate a
gene’s expression because it affects the accessibil-
ity of transcription factors and RNA Polymerase
to the gene region. When the genomic DNA is
wound tightly in a nucleosome, transcriptional
elements may have a difficult time “finding” the
gene they intend to activate. In order to provide
easier access to particular genes, the chromatin
structure will temporarily reorganize through a
process known as chromatin remodeling. The
biochemistry behind chromatin remodeling is
quite complex; however, we can generalize by
understanding that a group of special remodeling
proteins will shift the nucleosome cores along the
genomic DNA strand to increase access to par-
ticular regions of DNA.
• Chromatin modification: Chromatin can also be
modified by chemical modifications known as
methylation and acetylation, where methyl or
acetyl chemical groups are added to the nucleo-
some respectively. These chemical modifications
typically occur at lysine residues along the nucle-
osome protein and they can affect how the tran-
scriptional machinery is able to access and bind
to regions of genomic DNA.
• microRNAs (miRNA): Micro RNAs (miRNAs)
are an interesting class of small RNA molecules
(22 nucleotides long on average) that are now
known to post-transcriptionally regulate a large
number of mammalian genes. It is thought that
they primarily act to silence genes by binding to
the 3’ untranslated region (UTR) or their mRNA
transcripts. The human genome is predicted to
express ~1000 miRNAs which are thought to be
differentially expressed across different tissues
A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 13
and cell types. Many miRNAs are found in the
intergenic regions of the human genome, and
hence offer a functional basis for at least a por-
tion of the 98% of non-coding human DNA.
1.5 The human epigenome
Just when geneticists and genomicists were getting
comfortable with the recently sequenced human
genome in the early 2000s, with high hopes that it
would help to unravel the myriad genetic mysteries
behind human nature and disease quickly, several
breakthrough studies demonstrated that genes could
be modified in ways that could alter their gene func-
tion without changing the DNA sequence. Further-
more, it was revealed that these modifications that
did not alter the DNA sequence were heritable from
parent to offspring. This phenomenon, known as
epigenetics, clearly presented a challenge to the clas-
sical dogmas of molecular biology and genetic inher-
itance which held DNA is the source of genetic
inheritance between generations, and thus compli-
cated our view of genomics even further. The term
epigenetics simply refers to any heritable changes
that happen above the level of DNA sequence. There
are several complex processes that can facilitate epi-
genetic modifications, several of which we have
already described, such as methylation. The effects
of epigenetics are so widespread and important to
normal cellular functioning as well as human health
and disease that researchers now often refer to the
human epigenome as an entity that is completely
distinct from the human genome. Although this dis-
tinction is fuzzy, there is currently a Human Epige-
nome Project underway (<http://www.epigenome.
org>) that aims to identify and catalog all epigenetic
features across the human genome in similar spirit
to how the human genome was initially decoded.
Since researchers still have much to learn about the
human epigenome, it will not be discussed heavily
in the first edition of this book. However, you should
understand at least a few important implications of
epigenomics to prevent you from thinking that DNA
is the only way to explain variability and inheritance
of traits.
• Genomic imprinting: Under the classic model of
genetic inheritance, you inherit one copy of a gene
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
14 EXPLORING PERSONAL GENOMICS
from your mother, and another copy of the same
gene from your father, and both of these genes
will be “turned on” (i.e. expressed) in your cells
with equal probability. However, in some cases
you will inherit genes in this way and one of the
genes, either the paternal or maternal copy, will
be specifically inactivated. Therefore despite the
fact that you have two copies of a gene, you will
express the gene in a parent-of-origin-specific
manner. This phenomenon is called genomic
imprinting, and although it affects a relatively
small percentage of genes in the human genome,
it can have substantial consequences for an indi-
vidual’s development and disease risk. A gene
will become imprinted, or “stamped”, in either
the sperm or the egg through epigenetic mech-
anisms (typically methylation). This imprint-
ing often happens in the promoter region of a
gene and permanently prevents the gene from
being expressed in the offspring. Rarely, errone-
ously imprinted copies of the same gene can be
inherited from both the father and the mother,
resulting in no expressed copies in the offspring.
This phenomenon can result in disease, such as
the developmental disorders Prader-Willi syn-
drome and Angelman syndrome.
• Environmental factors: Understanding that dis-
ease is the product of genes and environment,
the discovery of epigenetics led researchers to
study possible interactions between the environ-
ment and epigenetic modification. It turns out
that the environment can indeed work through
epigenetic mechanisms to alter our biology in
profound ways. A flurry of work has linked
environmental pollutants to epigenetic modifi-
cations in genomic DNA that can, for example,
modify genes in a developing fetus resulting in
the child developing asthma in childhood, or
cause epigenetic modifications to somatic cells
that lead to dramatically increased cancer risk.
It has also been shown, most strongly in rats and
mice, that environmental pollution can cause
epigenetic modifications to germ-line DNA
(i.e. in sperm and eggs) and be passed down
through several generations of offspring. There-
fore environmental context can be inherited
across generations through epigenetic imprint-
ing of DNA!
1.6 Replication and reproduction
The molecular and cellular biology behind cellular
replication and human reproduction are now regu-
lar parts of the curriculum of secondary school edu-
cation in many countries. Therefore, we will only
aim to refresh the reader’s memory a handful of key
concepts pertaining to these phenomena that are
important for understanding other topics discussed
in this book.
Foremost, it is important to understand the dis-
tinction between the two forms of cellular division
and replication that happen in the human body. In
mitosis, a cell will divide and replicate the nuclear
DNA such that two diploid copies of a cell are pro-
duced. This form of cell replication occurs in regen-
erating body tissues such that tissues can grow or
renew, such as in muscle repair after exercise. Any
changes that occur to the DNA through errors in
mitosis are called somatic mutations, which are not
passed on to offspring because genetic material
from somatic cells (i.e. cells that make up your body)
are not passed to offspring during sexual reproduc-
tion. In meiosis, diploid germ cells will recombine
DNA and divide to produce four haploid gamete
cells (i.e. sperm or egg), which contain the genetic
material passed to offspring through sexual repro-
duction. Modifications to DNA during meiosis can
have profound impact on the offspring, but do not
affect the “host” body in any way. Although these
terms sound similar, you can see that they are very
different, and therefore it is important to distinguish
them. One mnemonic that is useful in this case is to
remember the phrase “mitosis happens in my toe” to
recall that mitosis occurs in somatic cells and meiosis
therefore happens in reproductive cells.
An important event that occurs during meiosis is
homologous recombination, which works to main-
tain a sufficient level of genetic variation within a
population. Homologous recombination, often
referred to as simply “recombination”, is one of the
reasons why a set of siblings from the same parents
will have different characteristics despite their
shared genetic origins. The biology underlying
meiosis and recombination is quite complex, but all
you should understand is that at some point during
meiosis, homologous chromosomes in the germ
cell will pair up and exchange DNA segments
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 15

Paternal Maternal
Chromosome Chromosome

DNA Replication Crossing Over

Recombination berween two homologous Chromosomes

Paternal Maternal Crossing Over Recombined

Chromosomes

Figure 1.10 Recombination: During meiosis, genetic material is exchanged between homologous pairs of chromosomes in a process known as
recombination. While the resulting genetic material is approximately the same in the end, the specific combination of variants may be shuffled
around. Image adapted from the National Human Genome Research Institute.

which recombine to form new chromosomal vari-
ants. More plainly, let us assume we are talking
about meiosis in the father’s germ cell. The father
inherited two copies of a particular chromosome
(for instance, chromosome 6): one from his mother
and another from his father. At some point during
meiosis, these maternal and paternal copies of
chromosome 6 will pair up and exchange segments
of DNA through a process called “crossover” (Fig-
ure 1.10). One analogy is to think of each of the
parental chromosomes as a strip of colored paper,
one blue and the other pink. Then, imagine cutting
these strips at random points, switching the blue
and pink segments, and gluing them together.
Although the two new strips contain the same total
“information” as before, they have been recom-
bined to create two novel variants of this informa-
tion. The points at which chromosomes break and
interchange with their homologous partner are not
completely random: regions that undergo recombi-
nation more frequently are known as recombina-
tion “hotspots”.
As you might imagine, this pseudo-random shuf-
fling of DNA between ancestral paternal and mater-
nal lineages provides the genetic variability that is
necessary to allow future generations to adapt to a
constantly changing environment. However, the
stochastic (i.e. random) nature of recombination
also provides the opportunity for deleterious genetic
events to occur. For example, “errors” during recom-
bination can lead to chromosomal variants that have
extra copies of genes, or might even have entire
genes completely deleted from the genomic DNA.
Such changes in gene number are called Copy
Number Variants (CNV) and while they might be
advantageous given the right environment, they are
most often have either a neutral or detrimental
affect to the offspring.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
16 EXPLORING PERSONAL GENOMICS
1.7 Genetic variation
Genetic variation is simply a catch-phrase to
describe all the observed variability in genetic and
genomic characteristics between individuals and
populations. In genetic variation lie the keys to
understanding differences in traits, such as height
or eye color, as well as differences in dispositions to
genetic or complex diseases, and genetic markers of
heredity and ancestry. Current estimates put the
genetic similarity between any two individual
humans at approximately 99.9% on average. A 0.1%
difference might not seem like a big deal, but
remember that the human genome comprises just
more than 3 billion base pairs, so a 0.1% difference
entails around 3,000,000 base-pair differences!
Many of these differences are likely to be benign,
occurring in inconsequential regions of non-
functional DNA, or contributing to common differ-
ences in human traits, such as height or hair color.
Yet some proportion of this variability will cause
individuals to harbor recessive genetic diseases, be
more or less susceptible to complex disease, or cause
them to metabolize medications differently. There-
fore, relatively small changes across the vast human
genome are the focal point of many modern genomic
studies. We learn more about genetic variation in
the human population in later chapters on ancestry
and trait associations.
1.7.1 Polymorphism
Polymorphism is a word you will hear frequently
when you read scientific articles pertaining to
genomics. If you consider its Greek roots, the term
polymorphism simply means “many forms”. In the
context of genetics and genomics, it is used to
describe the genetic variability that can be found
along the human genome (Figure 1.11). A specific
location on the genome is called a locus, and there-
fore, we often say that a gene is found at a particular
locus. If we were to sequence a gene in a number if
unrelated individuals, we might find small differ-
ences among individuals in the nucleotide sequence
used to encode the gene. Therefore, we can say that
several forms, or alleles, exist for this gene in the
population, and that the gene is polymorphic. Let
us consider just a single position in one of the cod-
ing region (exon) of the gene’s nucleotide sequence.
If we sample the DNA of the population we might
find that 80% of the population has an adenine (A)
at this position, while the remaining 20% has a
guanine (G) at this position. We would call this a
single nucleotide polymorphism (SNP), pronounced
“snip”, where A is the major allele and G is the
minor allele, based on the population frequency.
SNPs are the most abundant polymorphisms found
in the human genome (Figure 1.11).
It is important to understand that all of the alleles
at a position can be “healthy”, in that the major alle-
les are not necessarily biologically preferred or nec-
essary for maintaining proper function. Individuals
with the minor allele can be just as biologically
robust as those carrying the major allele. The SNP
may simply be not functionally relevant, or for cod-
ing regions, recall the earlier discussion about the
human codon table and how redundancy is built
into the genetic code. Since SNPs are known to be
associated with disease, it is useful to characterize
SNPs across the human genome. The government
funded HapMap project (<http://hapmap.ncbi.
nlm.nih.gov/>) set out to catalog all of the “normal”
variation found among human populations by
identifying a large number of SNPs across globally
and ethnically distributed human populations. As
we will find out first-hand in later chapters, the
HapMap project is a highly valuable tool for explor-
ing personal genomics.
Person 1 ATCGCATCGAT ACGCTACGCTACG
Person 2 ATCGCATCGAT GCGCTACGCTACG
Person 3 ATCGCATCGAT GCGCTACGCTACG
Person 4 ATCGCATCGAT ACGCTACGCTACG
Person 5 ATCGCATCGAT ACGCTACGCTACG
Single Nucleotide Polymorphism (SNP)
Figure 1.11 Single Nucleotide Polymorphisms: A single base pair
mutation (point mutation) that is observed in more than one percent
of the population is known as a single nucleotide polymorphism
(SNP). These mutations, which most often only have two alleles, arose
at some point in human history and have spread throughout human
populations. In this example, we observe a SNP with two alleles
(A and G).
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 17

Normal
Amino Acids Ala Ile Arg Leu Gly Tyr Ser Ala Cys Ile His Val Ala Ile Arg
tRNA
anticodon CGA UAUUCC GAUCCA AUG UCA CGU ACG UAU GUG CAU CGA UAU GCG Protein
mRNA GCU AUA AGG CUA GGU UAC AGU GCA UGC AUA CAC GUA GCU AUA CGC
5' codons 3'

Missense mutation
Amino Acids Ala Ile Arg Leu Ala Tyr Ser Ala Cys Ile His Val Ala Ile Arg
tRNA
anticodon CGA UAU UCC GAU CGA AUGUCA CGUACG UAU GUGCAU CGAUAU GCG
mRNA GCU AUA AGG CUA GCU UAC AGU GCA UGC AUA CAC GUA GCU AUA CGC Protein
5' codons 3`
Missense mutation

Nonsense mutation
Amino Acids Ala Ile Arg Leu Gly Tyr Ser Ala Cys stop
tRNA Protein
anticodon CGA UAU UCC GAU CCA AUG UCA CGU ACG AUU
mRNA GCU AUA AGG CUA GGU UAC AGU GCA UGC UAA CACGUAGCUAUACGC 3`
5' codons
Nonsense mutation

Figure 1.12 Types of coding SNPs: Some SNPs that occur in coding regions can affect the amino acid sequence of the protein. These SNPs are
known as non-synonymous SNPs and can result in a change in amino acids (missense) or the introduction or loss of a stop codon (nonsense and
nonstop, respectively). Additionally, synonymous SNPs (not pictured in the figure) are SNPs that change the DNA sequence, but do not affect the
amino acid sequence (as the same amino acid is used due to redundancy in the genetic code; e.g. a point mutation that changes a codon from
GCC to GCA). Image adapted from the National Human Genome Research Institute.

Polymorphisms come into existence by a muta-
tion at a previous moment in evolutionary history
that persisted and spread throughout the population
(see Box 1.3). These mutations can be caused by any
number of factors, such as errors in DNA replication,
radiation, mutagenic toxins, structural methylation,
or viral activity. Mutations can involve regions of
DNA spanning one to many base pairs of DNA.
When a mutation involves a single base pair, it is
typically referred to as a point mutation. Mutations
in regions of the genome coding for genes are of par-
ticular interest because they have the potential to
change the composition, and therefore the function
of the protein product encoded by the gene. If a
mutation occurs in one of the codons of a gene cod-
ing region, it can have one of several possible effects
(Figure 1.12). A mutation in this codon might not
change the amino acid at all because it might simply
mutate one codon into another codon that codes for
the same amino acid; this type of mutation is called
a synonymous mutation. It is generally thought that
synonymous mutations are benign, but recent stud-
ies have demonstrated that synonymous mutations
can impact the structure of the mRNA or affect the
translational efficiency of the tRNA. A mutation that
causes a region to code for a different amino acid is
called a non-synonymous or missense mutation.
The severity of these mutations depends on several
factors, such as the physiochemical differences
between the original and mutant mutation. Another
type of non-synonymous mutation is a mutation
that causes an amino-acid coding codon to mutate
into a stop codon, which is typically referred to as a
nonsense mutation; these are often more detrimen-
tal as they can result in truncation of the protein
encoded by the gene.
Although SNPs are one of the most prominent
types of genetic polymorphism in the human
genome, they are far from the only types found (Fig-
ure 1.13). SNPs have stolen the stage because the first
phase of the personal and medical genomics revolu-
tion was enabled by a specific type of genotyping
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

18 EXPLORING PERSONAL GENOMICS

Single nucleotide Insertion and deletion Nucleotide repeat

polymorphism (SNP) polymorphism (indel) polymorphism

C A T G C C G C A C

G T A C G G C G T G

C C T G C A C G C A C A C A C

G G A C G T G C G T G T G T G

Copy number variation

Gene

Gene

Gene Gene

Gene Gene

Deletion Duplication

Figure 1.13 Other types of mutations and polymorphisms: In addition to single nucleotide polymorphisms, other types of variations can also
alter the DNA and cause a functional change. For instance, indels can form when bases can be inserted or deleted, while the tandem duplication of
short segments can produce nucleotide repeat polymorphisms. When these mutations occur in coding regions, they can cause a “frameshift” in
codons. Additionally, larger variants can delete or duplicate whole genes or large segments of DNA, which are known as copy number variants and
structural variants (see Chapter 10). Reproduced from Hingorani, A. D., Shah, T., Kumari, M., Sofat, R. & Smeeth, L. Translating genomics into
improved healthcare. BMJ 341, c5945 (2010) with permission from BMJ Publishing Group Ltd.

technology called SNP microarrays, or “SNP chips”. • Repeat polymorphisms: Repeat polymorphisms
With the advent of affordable full-genome sequenc- are repeated patterns of short DNA sequences (e.g.
ing technology, other forms of polymorphism will be GATAGATA) found in abundance in the human
easier to measure and evaluate. Important polymor- genome. Since the repeat sequences are typi-
phisms found in the human population include: cally short, ranging from two to five base pairs in
length, these repeats are often called short tandem
• Single nucleotide polymorphism (SNP): SNPs
repeats (STR). STR polymorphisms are impor-
entail a class of polymorphisms where a single
tant for forensics, because STR polymorphisms
nucleotide in the human genome is found to be
between individuals serve as the basis for the
different among members of a population.
DNA fingerprinting used by law enforcement
• Insertion/Deletion (indel): Indel polymor-
agencies. Another larger type of repeat polymor-
phisms represent a type of genetic variation in
phism, called Alu repeats, is also used frequently
which sections of nucleotide sequence are found
to study ancestry and population structure.
to be inserted into or deleted from a region in the
• Copy number variations (CNV) and struc-
human genome. Although indel polymorphisms
tural variants (SVs): CNVs describe segments
can be associated with disease, especially in
of DNA which have a different number of cop-
coding regions, there are many cases of benign
ies between individual genomes. Although this
indel polymorphisms in the population.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
Box 1.3 “Molecular Evolution”
Evolution is critical to understanding the origins of species,
genome architecture, and even human disease mutations.
Evolutionary studies comparing genomes across species
have contributed significantly to our understanding of the
human genome. Perhaps one of the most critical things to
understand is that evolutionary forces are the primary drivers
of genetic variability both across species and within popula-
tions. The evolutionary forces acting on the human genome
have contributed to substantial changes in our species, even
within the past several thousand years. Although evolution is
typically discussed in terms of millions of years, it is an impor-
tant note that evolution is ongoing, and that evolutionary
forces come to bear on each human generation born into
existence today. Additionally, the environment is very impor-
tant in shaping evolution and our interactions with other
species, such as the bacteria in our digestive tracts, can have
a profound affect on our evolution. For example, it appears
that up to 8% of the human genome is of viral origin, mean-
ing that transposable DNA from ancient viruses was inte-
grated into our ancestral genome some time in the past and
has now become part of the blueprint for who we are as a
species! A number of evolutionary forces have shaped the
human genome and are important for understanding how
genomes operate in modern humans.
• Gene Duplication: Gene duplication is one of the pri-
• Mutation: A mutation is simply a change to the DNA
sequence of a genome. With regards to evolution, we are
primarily concerned with mutations that occur in the germ
line of an organism as these mutations can be passed on
to offspring.
• Natural selection: Natural selection is a key principal
in molecular evolution, which pertains to the increase or
decrease in particular genetic traits as a function of fitness
and reproductive success. Although natural selection is a
complex process, it can be thought of as a kind of filter
that removes suboptimal alleles from a population so that
the population is better adapted to its environment.
• Genetic drift: In each generation of a population, there
will be individuals who will leave behind more offspring
than others simply by chance. After several generations
the effects of this random sampling will cause the allele
frequencies to “drift” randomly towards higher or lower
frequencies, which is called genetic drift (Figure 1.14).
A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 19
An important thing to understand about genetic drift is
that if we observe changes in allele frequencies within
a population, the observed changes are not necessarily
the result of natural selection, but rather could simply be
explained by random genetic drift. In a sufficiently large
population, allele frequencies will tend to drift around
some stable population equilibrium (see Box 1.4). The
effects of genetic drift are more profound in smaller
populations, such as populations that experience a pop-
ulation bottleneck (Figure 1.15). One important con-
sequence of this is called the founder effect. Imagine
a fictional ancient human population of 100 individuals
that decided to separate from their main population and
establish a new, isolated human settlement in some fara-
way land. The reduced genetic diversity and smaller pool
of reproducing individuals in this population could cause
extreme, random shifts in allele frequencies for this popu-
lation as it expands. We might observe that formerly rare
alleles reach fixation at 100% frequency in the popula-
tion in just a few generations. Such populations can be
more prone to certain recessive genetic disorders due to
the reduced genetic diversity in their populations.
mary mechanisms for functional innovation in genome
evolution. Many of the important gene families in our
genome, such as hormones and cell surface receptors
(which facilitate the effects of medicinal drugs) came
about through gene duplication events in evolution-
ary history. Gene duplication occurs when any region
of genomic DNA containing a gene is duplicated and
becomes fixed in the germ line of a species. Duplications
can occur as the result of errors during recombination,
the activity of retrotransposons, or in some cases entire
chromosomes can be duplicated. From an evolutionary
standpoint duplications are very interesting because a
duplicated “copy” of a gene often has reduced selective
pressures acting against it because it is usually function-
ally redundant to the original copy it was duplicated
from. This means that it is more free to acquire mutations
and potentially evolve its function towards the formation
of a completely novel gene.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi
20 EXPLORING PERSONAL GENOMICS
Box 1.4 Hardy-Weinberg equilibrium
Developed independently by British mathematician G.H.
Hardy and German physician Wilhelm Weinberg, the Hardy-
Weinberg Equilibrium (HWE) model is a theoretical mathe-
matical model concerning the probability and distribution
genotype allele frequencies in a population. Somewhat anal-
ogous to rules of Mendelian inheritance that describe the
transmission of alleles at the family level, the HWE estab-
lishes a framework that can be used to model and predict
genotype frequencies in large, stable populations. To illus-
trate, let’s consider a single SNP locus that has three possible
states: homozygous for the minor allele (aa), heterozygous
(Aa), or homozygous for the major allele (AA). The parame-
ters of the HWE equations are the frequency of the major
allele, denoted p, and the frequency of the minor allele,
denoted q, with the relationships between q and p expressed
in the equilibrium model expressed as
p2 + 2pq+ q2 = 1
Note that because we are dealing in allele frequencies, we
can easily infer the frequency of one allele given the fre-
quency of the other. For example, if we measure a SNP in a
population and find that 90% of the individuals carry the
major allele (A), then we can subtract this proportion from
one to determine that 10% of individuals in the population
carry the minor allele (a). We can then use the HWE equation
to estimate the frequency of heterozygotes (2pq) in the
population.
0.92 + 2pq + 0.12 =1
0.81+ 2pq + 0.01=1
2pq =1- 0.82 = 0.18
can pertain to any segment of DNA ranging
from a few kilobases to even megabases in size,
it is easiest to conceptualize CNVs in terms of
gene regions. Since humans are diploid, they
typically inherit two copies of any one gene;
one from your father and the other from your
mother. However, an individual may only have
one copy of a gene because the other copy was
deleted in one of the parental genomes, or have
three copies of a gene because the gene was
duplicated. CNVs can be inherited or can hap-
The HWE would assert that the frequencies and relative
proportions of these genotypes would remain stable (i.e. in
equilibrium) over time if all the assumptions of the HWE are
satisfied. The assumptions of the HWE, all of which must be
satisfied to assert HWE, are:
• The population is (infinitely) large
• The mating patterns between population members are
random
• All members of the population have equal reproductive
success
• Males and females have similar allele frequencies (more
likely on an autosomal locus)
• Mutation is not occurring
• There is no significant migration in or out of the
population
• All genotypes have equal fitness (i.e. there is no selection)
At this point in the chapter, even a genomics neophyte
should carry enough understanding to suspect that the
assumptions of HWE are likely to be violated in the majority
of cases. In fact, this is why HWE is useful, because deviation
from HWE is often suggestive that the locus has been
affected by non-equilibrium forces such as mutation or evo-
lutionary selection. One of the most straightforward means
to statistically evaluate deviations from HWE is to perform a
Pearson’s chi-squared test to look for significant deviation
between the observed genotype frequencies, compared to
the genotype frequencies that would expected under HWE.
It is also reasonable to use HWE as a basis for inferring popu-
lation genotype frequencies where only the frequency of a
single allele for a locus is known or reported by a genetic
association study, for example.
pen spontaneously, the latter of which is known
as de novo CNVs, due to errors in the DNA rep-
lication machinery. CNVs are implicated in a
number of diseases, such as schizophrenia, and
are suspected to underlie many more genetic
disorders. Structural variants are rearrange-
ments of DNA that can affect large regions of
DNA, similar to CNVs, including insertions,
deletions, and inversions that can affect entire
genes. We will discuss CNVs and SVs further in
Chapter 11.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 21

100
Population size (N) = 50
90
80
Allele Frequency (%)

70
60
50
40
30
20
10
0
0 10 20 30 40 50
Generations

100
Population size (N) = 5,000
90
80
Allele Frequency (%)

70 Original Bottleneck Next generation

60
50
40
30
20
10
0
0 10 20 30
Generations
Figure 1.14 Effects of genetic drift on allele frequencies: These
graphs show the effect of genetic drift on allele frequencies in a small
(top; N = 50) and large (bottom; N = 5000) human population for
three distinct SNP variants exhibiting Hardy-Weinberg equilibrium. In
this example, each allele begins with a population frequency of 50%,
and the frequency begins to vary over successive populations due to
the effects of genetic drift. The effects of random sampling error are
much more drastic in small population, causing one of the alleles to
drift towards fixation (solid line), and another towards elimination (i.e.
loss of the allele from the population) (dashed line). In the larger
population, the effects of genetic drift due to sampling error are less
drastic, and the allele frequencies do not vary far from their original
population frequency over successive generations.
1.7.2 Linkage disequilibrium
Thus far, we’ve discussed the different types of
mutations independently, where a single mutation
occurs and may spread throughout the population
until it reaches high enough frequency to become a
polymorphism. However, the mechanics of genomic
inheritance implies that the inheritance of any two
polymorphisms may not be independent. To illus-
trate how this may come about, consider two germ-
line point mutations that occur on the same copy of
a chromosome 100 bases apart. When a chromo-
some is transmitted to a sperm or an egg cell during
population event after bottleneck
Figure 1.15 Population bottleneck: This phrase describes a scenario
in which a population’s size is substantially reduced for at least one
generation, perhaps due to a pandemic disease, war, natural disaster,
or some other event. The reduced population will reduce the amount
of total genetic variation in the population, and, as illustrated in
40 50 Figure 1.14, the effects of genetic drift on allele frequencies can be
much more drastic in smaller populations. In this illustration, insects of
the same species are contained within a bottle, and are distinguished
by their external color, which is controlled by a single allele. Even if a
population quickly recovers back to a large population size in
subsequent generations, the effects of the population bottleneck, in
terms of reduced genetic variation, can be apparent in the genomic
makeup of a population. The genomic patterns of modern populations
of European ancestry seem to bear hallmarks of a historical
population bottleneck, marked by reduced overall patterns of
heterozygosity compared to other worldwide populations.
meiosis (see Section 1.6—Replication and reproduc-
tion), these two neighboring SNPs will then be
transmitted together, unless there is a recombina-
tion event in between. Furthermore, the probability
of recombination is relatively low for small sections
(such as the two SNPs 100 base pairs apart): there
are, on average, 35 recombination events per meio-
sis, which corresponds to roughly one recombina-
tion every 100 Mb. Over time, this process continues
and leads to the correlated inheritance of SNPs,
which are known as linked SNPs (Figure 1.16). More
broadly, this phenomenon is also termed linkage
disequilibrium (LD); when two SNPs are inherited
randomly (unlinked), they are said to be in equilib-
rium. Thus, a high level of LD implies that two SNPs
are “linked” and the measure of linkage is known as
R2, which is the level of correlation between the
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

22 EXPLORING PERSONAL GENOMICS

A C T
A C T
A C A
A T T
A C A
G T A
G T T
G T A
G T T
G T A

Linked Unlinked

Linked Unlinked
positions positions

Over
Time
Figure 1.16 Polymorphisms are often inherited together due to
linkage disequilibrium: As mutations arise, they are inherited together
with nearby alleles, unless a recombination event occurs in between.
Over time, as these alleles spread throughout the population, the
presence of one of the alleles is correlated or “linked” with the other.
These SNPs are said to be in linkage disequilibrium (LD).
Figure 1.17 Haplotype blocks: The phenomenon of linkage
disequilibrium (LD) often results in the correlated inheritance of
nearby variants. In this schematic, a stretch of DNA with SNPs
(represented by stars) is shown. Below is a commonly used display of
linkage disequilibrium patterns: darker shaded regions correspond to
more tightly linked SNPs. In this example, the white star SNP is
unlinked to the gray star SNPs, which are all linked with each other. To
read the diagram, follow down the diagonal from a SNP until the
opposite diagonal of the other SNP of interest. Here, because the
triangle is dark between the two marked SNPs, the two SNPs are in
high LD.

lotype blocks will be used in the design of genotype-

SNPs. A perfect correction (fully linked SNPs, where
genotype A of SNP 1 is always observed with geno-
type B of SNP 2) occurs when R2 = 1.0, and two ran-
dom SNPs will have R2 = 0.
As we progress through our discussion of per-
sonal genomics, the concept of LD will become cru-
cial for clinical and phenotype risk analysis, as well
as ancestry analysis. One major reason for this is the
development of haplotype blocks in a population
(Figure 1.17). These blocks are segments of SNPs
that are often inherited as a group; thus, we would
only need to measure one SNP (a tag SNP) to pre-
dict the genotype of another (in a process known as
imputation). In reality, the situation is not always as
clear-cut: unless we can find a SNP in perfect LD
(R2 = 1), we may not be able to fully and accurately
impute our SNP of interest. Additionally, different
populations will have different patterns of linkage
disequilibrium: populations that have undergone
bottlenecks typically have longer haplotype blocks
than those that have not. As we will see, these hap-
phenotype association experiments (Chapter 2) and
the application of these associations to our personal
genomes (Chapter 6), as well as ancestry analysis
(Chapter 5).
Further reading
Unfortunately, there is more fascinating breadth and depth
to genomics that could not be covered by this primer.
We suggest the following materials and resources for
those inclined to expand their study and understanding
of molecular biology and genomics.
Alberts, B. (2008) Molecular biology of the cell. New York,
NY: Garland Pub.
Berger, S. L. (2000) Gene regulation. Local or global?
Nature 408, 412–13, 415.
Brown, T. A. (2007) Genomes Three. London: Taylor
Francis.
Chen, K. & Rajewsky, N. (2007) The evolution of gene reg-
ulation by transcription factors and microRNAs. Nature
Reviews Genetics 8, 93–103.
Collins, F. S. (2011) The Language of Life. New York, NY:
Harper Perennial.
 OUP CORRECTED PROOF – FINAL, 11/29/2012, SPi

A G E N T L E I N T R O D U C T I O N TO G E N O M I C S 23

Hingorani, A. D., Shah, T., Kumari, M., Sofat, R. & Smeeth,
L. (2010) Translating genomics into improved health-
care. BMJ 341, c5945.
Lynch, M. (2007) The origins of genome architecture. Sunder-
land, MA: Sinauer Associates Inc).
Nei, M. & Kumar, S. (2000) Molecular evolution and phyloge-
netics. New York: Oxford University Press.
Ridley, M. (2006) Genome. New York, NY: Harper Perennial.
Watson, J. D., Caudy, A. A., Myers, R. & Witkowski, J. A.
(2007) Recombinant DNA. W. H. Freeman.