Report

Modeling the 3D Geometry of the Cortical Surface

with Genetic Ancestry
Graphical Abstract Authors
Chun Chieh Fan, Hauke Bartsch,
Andrew J. Schork, ..., Nicholas
J. Schork, Terry L. Jernigan,
Anders M. Dale

Correspondence
amdale@ucsd.edu

In Brief
Fan et al. show that human cortical
surface robustly predicts an individual’s
genetic ancestry despite that populations
have been shaped by waves of migrations
and admixture events. For each
continental ancestry, the regional
patterns of cortical folding and
gyrification are unique and complex.

Highlights
d Geometry of the human cortical surface contains rich
ancestral information

d The most informative features are regional patterns of cortical

folding and gyrification

d This study provides insight on the influence of population

structure on brain shape

Fan et al., 2015, Current Biology 25, 1988–1992

August 3, 2015 ª2015 Elsevier Ltd All rights reserved
http://dx.doi.org/10.1016/j.cub.2015.06.006
 Current Biology

Report

Modeling the 3D Geometry of the Cortical Surface

with Genetic Ancestry
Chun Chieh Fan,1 Hauke Bartsch,2 Andrew J. Schork,1 Chi-Hua Chen,3 Yunpeng Wang,2,4,5 Min-Tzu Lo,3
Timothy T. Brown,2,5 Joshua M. Kuperman,2,3 Donald J. Hagler, Jr.,2,3 Nicholas J. Schork,6 Terry L. Jernigan,1,3,7,8
Anders M. Dale,1,2,3,5,8,* and the Pediatric Imaging, Neurocognition, and Genetics Study
1Department of Cognitive Science, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA
2Multimodal Imaging Laboratory, University of California, San Diego School of Medicine, 9500 Gilman Drive, La Jolla, CA 92037, USA
3Department of Radiology, University of California, San Diego School of Medicine, 9500 Gilman Drive, La Jolla, CA 92037, USA
4Norwegian Centre for Mental Disorders Research (NORMENT), KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine,

University of Oslo, 0424 Oslo, Norway

5Department of Neuroscience, University of California, San Diego School of Medicine, 9500 Gilman Drive, La Jolla, CA 92037, USA
6J. Craig Venter Institute, Capricorn Lane, La Jolla, CA 92037, USA
7Center for Human Development, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92161, USA
8Department of Psychiatry, University of California, San Diego School of Medicine, 9500 Gilman Drive, La Jolla, CA 92037, USA

*Correspondence: amdale@ucsd.edu
http://dx.doi.org/10.1016/j.cub.2015.06.006

SUMMARY RESULTS

Knowing how the human brain is shaped by migra-
tion and admixture is a critical step in studying
human evolution [1, 2], as well as in preventing the
bias of hidden population structure in brain research
[3, 4]. Yet, the neuroanatomical differences engen-
dered by population history are still poorly under-
stood. Most of the inference relies on craniometric
measurements, because morphology of the brain is
presumed to be the neurocranium’s main shaping
force before bones are fused and ossified [5].
Although studies have shown that the shape varia-
tions of cranial bones are consistent with popula-
tion history [6–8], it is unknown how much human
ancestry information is retained by the human
cortical surface. In our group’s previous study, we
found that area measures of cortical surface and total
brain volumes of individuals of European descent in
the United States correlate significantly with their
ancestral geographic locations in Europe [9]. Here,
we demonstrate that the three-dimensional geome-
try of cortical surface is highly predictive of individ-
uals’ genetic ancestry in West Africa, Europe, East
Asia, and America, even though their genetic back-
ground has been shaped by multiple waves of migra-
tory and admixture events. The geometry of the
cortical surface contains richer information about
ancestry than the areal variability of the cortical
surface, independent of total brain volumes. Be-
sides explaining more ancestry variance than other
brain imaging measurements, the 3D geometry of
the cortical surface further characterizes distinct
regional patterns in the folding and gyrification of
the human brain associated with each ancestral
lineage.
The participants were recruited as part of the Pediatric Imaging,
Neurocognition, and Genetics (PING) study. A detailed overview
of the study can be found in previous publications (e.g., [3, 4, 10]),
and research protocols and data are publicly available online [11].
Briefly, PING was a multisite project recruiting children and ado-
lescents from ages 3 to 21 at ten sites in the United States. All
participants were screened for history of major developmental,
psychiatric, and neurological disorders; brain injury; and other
medical conditions that affect development. Participants then
received neurodevelopmental assessments, standardized multi-
modal neuroimaging, and genome-wide genotyping. The overall
PING sample consisted of 1,493 participants; 1,152 individuals
remained after quality control of the genotyping and neuroimag-
ing data (for quality-control processes and demographics of the
participants, see Supplemental Experimental Procedures and
Table S1). We focused our analyses on 562 individuals older
than 12 years (289 males, mean age 16.6 years, standard devia-
tion 2.6 years). Considering that the morphological features of
cortical surface change little after age 12 [10], this stratified
approach further reduced the residual confounds of develop-
mental effects.
The proportions of genetic ancestry were estimated using
principal component (PC) analysis with whole-genome SNP
reference panels for ancestry [12–14]. Four continental popula-
tions were used as ancestral references: West Africa (YRI,
Yoruba in Ibadan), Europe (CEU, Utah residents with Northern
and Western European ancestry), East Asia (EA), and America
(NA, Native American). The metrics for summarizing genetic
ancestry in each ancestral component were standardized as
proportions ranging from 0% to 100%. These proportions repre-
sent how genetically similar an individual is to the reference pop-
ulation [14].
Morphological Prediction for Genetic Ancestry
We first tested whether the surface geometry of the cerebral cor-
tex predicted the proportion of genetic ancestry among partici-
pants. To characterize variation in the geometry, we reconstructed

1988 Current Biology 25, 1988–1992, August 3, 2015 ª2015 Elsevier Ltd All rights reserved

the cortical surfaces from all individuals’ T1-weighted scans and
then represented the positions of the corresponding surface
vertices using standard 3D Cartesian coordinates. The recon-
struction and registration processes ensure that each vertex on
the reconstructed cortical surface is located in a homologous po-
sition with respect to the curvature patterns for individuals [15, 16].
Taking the coordinates of all vertices as a whole, we then have
information about shape variation of the cortical surface, including
aspect ratios, sulcal depth, and gyrification. The prediction
models were fit with ridge regression while treating gender, age,
age squared, total brain volumes, and the scanner on which the
image data were acquired as nuisance covariates. The model per-
formance was evaluated using leave-one-out cross-validation
(LOOCV).
As Figure 1 shows, the geometry of the cortical surface has
good predictive value for each of the ancestry components.
The variances explained by the models are 66% for ancestry
in YRI, 55% for ancestry in CEU, 49% for ancestry in EA,
and 47% for ancestry in NA. To determine to what degree
the geometric differences reflect variation in area expansion
of cortical surface, comparable models were computed using
vertex-wise surface area (Table 1). Also, to examine possible
roles in the prediction of simpler morphological attributes,
such as aspect ratios of the cerebrum and volumes of subcor-
tical structures, we conducted comparable analyses predicting
ancestry from these measures. None had as much information
Current Biology 25, 1988–1992, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 1989
Figure 1. Predicting the Proportion of
Genetic Ancestry by Cortical Surface
Geometry
YRI: Yoruban, as a proxy for West African ancestry;
CEU: Utah residents with Northern and Western
European ancestry; EA: East Asian; NA: Native
American. In all predictive models, the variables
have been residualized with respect to age, age
squared, gender, total brain volumes, and scanner
used. All models excluded individuals with a 0%
proportion of genetic ancestry to that specific
component. The colors of the data points are
determined by the proportion of genetic ancestry
as illustrated in the key in the upper left panel.
LOOCV: leave-one-out cross-validation.
about ancestry as the geometry of
cortical surface did (Table 1).
Characterization of the Cortical
Shape Morphs
We then reconstructed the 3D geometry of
the cortical surface based on the linear
relationship we observed between cortical
surface geometry and proportion of ge-
netic ancestry. This allowed us to visualize
how the geometry of the cortical surface
changes as a function of increasing pro-
portion of genetic ancestry in each ances-
tral component. The morphing of 3D
cortical surfaces from neutral ancestry
(25% of genetic ancestry in all four compo-
nents) to 100% ancestry in each compo-
nent is demonstrated in Figure 2 (for dynamic morphing of surface
geometry, see Movies S1, S2, S3, and S4). As Figure 2 illustrates,
the textural contrasts between regions of the cortical surface indi-
cate that the morphing process has complex, unique patterns for
each ancestral component, while the intensity varies from region
to region. For example, as the proportion of the YRI component
increases, the temporal surfaces move posteriorly and inward.
The proportion of the CEU component is associated with protru-
sion of the occipital and frontal surfaces. Increases in the pro-
portion of the EA component are accompanied by variations in
temporal-parietal regions. The NA component is associated
with flattening of the frontal and occipital surfaces.
Figure 3 summarizes the mean magnitudes and variations of
the morphing in each cortical surface region defined by genetic
correlations [17]. The mean magnitudes vary from cortical region
to cortical region, corresponding to the description above. In
addition, YRI, EA, and NA all have relatively high magnitude
and variations of morphing in the posterolateral-temporal region.
DISCUSSION
Our data indicate that the unique folding patterns of gyri and
sulci are closely aligned with genetic ancestry. The geometry
robustly predicts each individual’s genetic background even
though the population has been shaped by waves of migration
and admixtures [12, 18]. A previous study, using only facial
Table 1. Percentage of Variance Explained in Different Predictive On the other hand, the global shapes of the reconstructed
Models cortical surface geometry match W.W. Howells’ description of
Cortical Surface Cortical Surface Brain Aspect Subcortical craniometry of 2,524 ancient human crania from 28 populations
Geometry Area Ratios Volumes [20]. Crania of African ancestry tended to have a narrower cranial
YRI 66% 17% 10% 5%
base, and those of Northern European ancestry had elongated oc-
cipital and frontal regions. Crania of East Asian ancestry had a high
CEU 55% 12% 2% 2%
cranial vault, and crania of Native American ancestry were flatter.
EA 49% 9% 6% 6%
Regarding the morphing differences of YRI, EA, and NA, all had
NA 47% 9% 9% 0% high magnitude and variations in the posterior-temporal regions
Cortical surface geometry and cortical surface area were sampled in (Figure 3).These findings are consistent with the notion that tem-
icosahedral level 4, which contains 642 vertices in each hemisphere. All poral bones contain more variations across ancestral groups [6].
models were fit with the same setting and evaluated with leave-one-out At first glance, these results are surprising because our model
cross-validation (LOOCV). Nuisance covariates gender, age, age
is based on the contemporary United States population, which
squared, total brain volumes, and scanner were regressed out before
is the historical product of migrations, slave trades, and local
calculating the variance explained in LOOCV.
admixture events [18, 21, 22]. Nevertheless, the coordinates of
reference-inferred PC space reflect information about individ-
features, achieved 64% explained variance in YRI ancestry uals’ ancestral origins (Figure S1) [14, 21, 23]. Our group’s previ-
among African Americans [19]. Our 3D representation of ous study also showed that individuals’ positions in PC space are
cortical surface geometry performs similarly in predicting YRI matched with their ancestral locations, rather than their current
ancestry and also performs well for the other three continental geographic locations [9]. Therefore, our 3D representation might
ancestries. As data in Table 1 show, the explanatory power to a certain degree reflect the neuroanatomical and/or neuro-
is not due to the differences in total brain volumes, nor to cranial changes along the human migratory path in the dispersal
the differences in areal expansion of the cortical surface. from Africa [24]. Based on our current model, we simulated what
Instead, regional folding patterns characterize each ancestral might be expected from the ‘‘out of Africa’’ scenario in the Sup-
lineage. plemental Experimental Procedures (Figure S3; Movie S5).

Figure 2. Color-Coded Morphing Process of

the 3D Geometry of the Cortical Surface
The still image illustrates how each vertex on the
cortical surface morphs from an ancestry-neutral
3D cortical surface (a 25% proportion of genetic
ancestry in all ancestral components) to a 3D
cortical surface with a 100% proportion of genetic
ancestry in a specific ancestral component. The
morphing coefficients were estimated from the
PING sample. Here, the colors represent the di-
rection of the morphing process. Movement along
the medial-lateral axis is coded in red, along
the anterior-posterior axis in green, and along the
dorsal-ventral axis in blue. The final color is the
combination of these three, depending on which
direction the vertices move. For each viewing
perspective, the coloring frame of reference is
rendered on the top of each column. The length of
each morphing line is the actual distance between
two 3D cortical surfaces. For dynamic morphing
animations, see Movies S1, S2, and S3.

1990 Current Biology 25, 1988–1992, August 3, 2015 ª2015 Elsevier Ltd All rights reserved

More precise characterization of an individual’s ancestral origins
would require more complex estimates of ancestry based on
global-scale reference panels [25]. Further understanding of
neuroanatomical change associated with the ‘‘out of Africa’’ sce-
nario based on brain imaging data will require future studies using
sampling methods similar to those of the Human Genome Diver-
sity Project [26].
It is important to note that these ancestry-related geometric fea-
tures of the cortical surface are not substantially attributable to
variation in cortical surface area. Previous studies of ancient crania
often interpreted the shape differences as evidence of relative size
alterations of different cortical functional domains [5, 27]. Our re-
sults suggest that in the case of the contemporary United States
population, the differences in cortical surface geometry might
not reflect variation in the relative surface area of different func-
tional cortical regions. In prior studies, regionalization of the cortex
Current Biology 25, 1988–1992, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 1991
Figure 3. Mean Magnitude and Variations of
Morphing across Twelve Regions of Cortical
Surface
The following regions are labeled at the top, as
defined in a previous publication [17]: 1, central re-
gion; 2, occipital cortex; 3, posterolateral temporal
region; 4, superior parietal region; 5, orbitofrontal
region; 6, superior temporal region; 7, inferior pari-
etal region; 8, dorsomedial frontal region; 9, ante-
romedial temporal region; 10, precuneus;11,
dorsolateral prefrontal cortex; 12, pars opercularis.
The Euclidean distances between cortical surface
of 100% ancestry and neutral ancestry were
calculated for each vertex. The mean and standard
deviations of the Euclidean distances for different
cortical regions are shown in the bar plots.
was linked to cognitive differences in
humans [3, 4]. Any functional significance
of the cortical surface geometry per se
remains to be established. The effects re-
ported here might be mediated by neutral
drift of the phenotypic variations [28].
They could also result from a complex
interaction between the brain and neuro-
cranium, with the former expanding while
the latter acts as physical resistance.
Nevertheless, the causal relationships be-
tween the observed shapes and crania
are beyond the scope of our current study.
An implication of our ancestry-related
3D models is that, unless properly
controlled for, hidden population struc-
tures could present a challenge in brain
imaging studies of admixed populations
[23]. The regional differences between
ancestral groups include changing sulcus
depths and folding angles. This issue be-
comes particularly relevant in large, multi-
site United States and international brain
imaging studies [29]. With the advent of
inexpensive high-throughput genotyping,
it is now possible to control for spurious
ancestry admixture effects by using genetically derived admix-
ture factors in the statistical analysis of data [3, 4]. It is also
possible that the phenomena we observed are linked with spe-
cific ancestral haplotypes. It may therefore be possible to use
the ancestral information to improve statistical power for gene
discovery with methods such as admixture mapping [30].
SUPPLEMENTAL INFORMATION
Supplemental Information includes three figures, one table, Supplemental
Experimental Procedures, and five movies and can be found with this article
online at http://dx.doi.org/10.1016/j.cub.2015.06.006.
ACKNOWLEDGMENTS
All data used in this article were obtained from the Pediatric Imaging, Neuro-
cognition, and Genetics (PING) database (http://pingstudy.ucsd.edu). The
investigators within PING contributed to the design and implementation of
PING but did not necessarily participate in this study. Data collection and
sharing for this project were funded by PING (NIH grant RC2DA029475). Fund-
ing was provided by the National Institute on Drug Abuse and the Eunice Ken-
nedy Shriver National Institute of Child Health and Human Development. PING
data are disseminated by the PING Coordinating Center at the Center for Hu-
man Development, University of California, San Diego. Other support was pro-
vided by National Institute of Mental Health grant R01MH100351. A.M.D. is a
founder of and holds equity interest in CorTechs Labs and serves on its scien-
tific advisory board. The terms of this arrangement have been reviewed and
approved by the University of California, San Diego, in accordance with its
Conflict of Interest policies.
Received: March 1, 2015
Revised: May 4, 2015
Accepted: June 1, 2015
Published: July 9, 2015
REFERENCES
1. Falk, D. (2014). Interpreting sulci on hominin endocasts: old hypotheses
and new findings. Front. Hum. Neurosci. 8, 134.
2. Holloway, R.L. (2008). The human brain evolving: a personal retrospective.
Front. Neuroanat. 37, 1–19.
3. Fjell, A.M., Walhovd, K.B., Brown, T.T., Kuperman, J.M., Chung, Y.,
Hagler, D.J., Jr., Venkatraman, V., Roddey, J.C., Erhart, M., McCabe,
C., et al.; Pediatric Imaging, Neurocognition, and Genetics Study (2012).
Multimodal imaging of the self-regulating developing brain. Proc. Natl.
Acad. Sci. USA 109, 19620–19625.
4. Walhovd, K.B., Fjell, A.M., Brown, T.T., Kuperman, J.M., Chung, Y.,
Hagler, D.J., Jr., Roddey, J.C., Erhart, M., McCabe, C., Akshoomoff, N.,
et al.; Pediatric Imaging, Neurocognition, and Genetics Study (2012).
Long-term influence of normal variation in neonatal characteristics on hu-
man brain development. Proc. Natl. Acad. Sci. USA 109, 20089–20094.
5. Bruner, E., de la Cuétara, J.M., Masters, M., Amano, H., and Ogihara, N.
(2014). Functional craniology and brain evolution: from paleontology to
biomedicine. Front. Neuroanat. 8, 19.
6. Reyes-Centeno, H., Ghirotto, S., Détroit, F., Grimaud-Hervé, D.,
Barbujani, G., and Harvati, K. (2014). Genomic and cranial phenotype
data support multiple modern human dispersals from Africa and a south-
ern route into Asia. Proc. Natl. Acad. Sci. USA 111, 7248–7253.
7. Manica, A., Amos, W., Balloux, F., and Hanihara, T. (2007). The effect of
ancient population bottlenecks on human phenotypic variation. Nature
448, 346–348.
8. Lieberman, D.E., McBratney, B.M., and Krovitz, G. (2002). The evolution
and development of cranial form in Homosapiens. Proc. Natl. Acad. Sci.
USA 99, 1134–1139.
9. Bakken, T.E., Dale, A.M., and Schork, N.J. (2011). A geographic cline of
skull and brain morphology among individuals of European Ancestry.
Hum. Hered. 72, 35–44.
10. Brown, T.T., Kuperman, J.M., Chung, Y., Erhart, M., McCabe, C., Hagler,
D.J., Jr., Venkatraman, V.K., Akshoomoff, N., Amaral, D.G., Bloss, C.S.,
et al. (2012). Neuroanatomical assessment of biological maturity. Curr.
Biol. 22, 1693–1698.
11. Jernigan, T.L., Brown, T.T., Hagler, D.J., Jr., Akshoomoff, N., Bartsch, H.,
Newman, E., Thompson, W.K., Bloss, C.S., Murray, S.S., Schork, N., et al.;
Pediatric Imaging, Neurocognition, and Genetics Study (2015). The
Pediatric Imaging, Neurocognition, and Genetics (PING) Data
Repository. Neuroimage. Published online May 1, 2015. http://dx.doi.
org/10.1016/j.neuroimage.2015.04.057.
1992 Current Biology 25, 1988–1992, August 3, 2015 ª2015 Elsevier Ltd All rights reserved
12. Reich, D., Patterson, N., Campbell, D., Tandon, A., Mazieres, S., Ray, N.,
Parra, M.V., Rojas, W., Duque, C., Mesa, N., et al. (2012). Reconstructing
Native American population history. Nature 488, 370–374.
13. Altshuler, D.M., Gibbs, R.A., Peltonen, L., Altshuler, D.M., Gibbs, R.A.,
Peltonen, L., Dermitzakis, E., Schaffner, S.F., Yu, F., Peltonen, L., et al.;
International HapMap 3 Consortium (2010). Integrating common and
rare genetic variation in diverse human populations. Nature 467, 52–58.
14. Chen, C.Y., Pollack, S., Hunter, D.J., Hirschhorn, J.N., Kraft, P., and Price,
A.L. (2013). Improved ancestry inference using weights from external
reference panels. Bioinformatics 29, 1399–1406.
15. Fischl, B., Sereno, M.I., and Dale, A.M. (1999). Cortical surface-based
analysis. II: Inflation, flattening, and a surface-based coordinate system.
Neuroimage 9, 195–207.
16. Dale, A.M., Fischl, B., and Sereno, M.I. (1999). Cortical surface-based
analysis. I. Segmentation and surface reconstruction. Neuroimage 9,
179–194.
17. Chen, C.H., Gutierrez, E.D., Thompson, W., Panizzon, M.S., Jernigan, T.L.,
Eyler, L.T., Fennema-Notestine, C., Jak, A.J., Neale, M.C., Franz, C.E.,
et al. (2012). Hierarchical genetic organization of human cortical surface
area. Science 335, 1634–1636.
18. Tishkoff, S.A., Reed, F.A., Friedlaender, F.R., Ehret, C., Ranciaro, A.,
Froment, A., Hirbo, J.B., Awomoyi, A.A., Bodo, J.M., Doumbo, O., et al.
(2009). The genetic structure and history of Africans and African
Americans. Science 324, 1035–1044.
19. Claes, P., Liberton, D.K., Daniels, K., Rosana, K.M., Quillen, E.E., Pearson,
L.N., McEvoy, B., Bauchet, M., Zaidi, A.A., Yao, W., et al. (2014). Modeling
3D facial shape from DNA. PLoS Genet. 10, e1004224.
20. Howells, W.W. (1989). Skull Shape and the Map: Craniometric Analyses in
the Dispersion of Modern Homo. (Harvard University Press), pp. 63–64.
21. Li, J.Z., Absher, D.M., Tang, H., Southwick, A.M., Casto, A.M.,
Ramachandran, S., Cann, H.M., Barsh, G.S., Feldman, M., Cavalli-
Sforza, L.L., and Myers, R.M. (2008). Worldwide human relationships in-
ferred from genome-wide patterns of variation. Science 319, 1100–1104.
22. Abdulla, M.A., Ahmed, I., Assawamakin, A., Bhak, J., Brahmachari, S.K.,
Calacal, G.C., Chaurasia, A., Chen, C.H., Chen, J., Chen, Y.T., et al.;
HUGO Pan-Asian SNP Consortium; Indian Genome Variation
Consortium (2009). Mapping human genetic diversity in Asia. Science
326, 1541–1545.
23. Price, A.L., Patterson, N.J., Plenge, R.M., Weinblatt, M.E., Shadick, N.A.,
and Reich, D. (2006). Principal components analysis corrects for stratifica-
tion in genome-wide association studies. Nat. Genet. 38, 904–909.
24. Cavalli-Sforza, L.L., Menozzi, P., and Piazza, A. (1993). Demic expansions
and human evolution. Science 259, 639–646.
25. Yang, W.Y., Novembre, J., Eskin, E., and Halperin, E. (2012). A model-
based approach for analysis of spatial structure in genetic data. Nat.
Genet. 44, 725–731.
26. Cavalli-Sforza, L.L. (2005). The Human Genome Diversity Project: past,
present and future. Nat. Rev. Genet. 6, 333–340.
27. Bruner, E., Manzi, G., and Arsuaga, J.L. (2003). Encephalization and allo-
metric trajectories in the genus Homo: evidence from the Neandertal and
modern lineages. Proc. Natl. Acad. Sci. USA 100, 15335–15340.
28. Roseman, C.C. (2004). Detecting interregionally diversifying natural selec-
tion on modern human cranial form by using matched molecular and
morphometric data. Proc. Natl. Acad. Sci. USA 101, 12824–12829.
29. Medland, S.E., Jahanshad, N., Neale, B.M., and Thompson, P.M. (2014).
Whole-genome analyses of whole-brain data: working within an expanded
search space. Nat. Neurosci. 17, 791–800.
30. Winkler, C.A., Nelson, G.W., and Smith, M.W. (2010). Admixture mapping
comes of age. Annu. Rev. Genomics Hum. Genet. 11, 65–89.
Current Biology
Supplemental Information

Modeling the 3D Geometry of the Cortical Surface

with Genetic Ancestry
Chun Chieh Fan, Hauke Bartsch, Andrew J. Schork, Chi-Hua Chen, Yunpeng Wang,
Min-Tzu Lo, Timothy T. Brown, Joshua M. Kuperman, Donald J. Hagler, Jr.,
Nicholas J. Schork, Terry L. Jernigan, Anders M. Dale, and the Pediatric Imaging,
Neurocognition, and Genetics Study
Figure S1. Distribution of the genetic ancestry. Related to Figure 1. A. Distribution of

the PING participants in the reference inferred PC space. B. Proportion of genetic

ancestry, transformed from the PC space. C. Consistency between proportion of genetic

ancestry and the model that mapped individual onto world geography [S12].
Figure S2. Quantile-Quantile plot of genetic ancestry by Gender. Related to Figure 1.

The distribution of genetic ancestry by gender is compared with the Q-Q plot. The overall

distributions are compatible with each other. Females tend to have more EA genetic

ancestry than males in some ranges of ancestral percentile while retaining the same

distribution at the tail and mid ranges.

Figure S3. Simulated neuroanatomical change along the path of dispersal from

Africa. Related to Figure 2. The color scheme is the same as in Figure 2. The migratory

path is represented as moving from Africa to Europe, then to East Asia, and finally to

America [S15].
Table S1. Demographic and covariates distributions of the PING participants
YRI CEU EA NA
Mean Proportion 14.6% 61.9% 18.5% 5%
Male % 51.3% 53.2% 47% 43.8%
Correlations with Age (r) -0.073 0.088 0.019 0.005
Correlations with 0.072 -0.032 -0.019 -0.049
Household Income
Scanners
1 21% 66% 11% 2%
2 22% 51% 12% 16%
3 5% 73% 17% 5%
4 30% 57% 8% 5%
5 17% 52% 12% 19%
6 7% 73% 16% 4%
7 7% 41% 50% 2%
8 8% 82% 8% 2%
9 7% 73% 12% 8%
10 18% 75% 6% 1%
11 19% 41% 4% 37%
12 31% 53% 11% 5%
Supplemental Experimental Procedures

Genotyping and ancestral inference

Saliva samples were processed by the PING Study Genomics Core at The Scripps

Translational Science Institute, where DNA was extracted and genotyped. The

genotyping was performed on the Illumina Human660W-Quad BeadChip with 550,000

single nucleotide polymorphism (SNP) markers. SNPs with genotyping call rates less

than 95%, minor allele frequencies less than 1%, or failing Hardy-Weinberg Equilibrium

(p < 10-7) are excluded. The final data have a total of 539,865 SNPs with an overall

genotyping rate of 99.95%. Individuals missing more than 5% SNPs were excluded from

the analyses.

Previous studies using PING data have shown that supervised clustering of genomic

data can group PING participants into continental populations, including African,

European, Asian, and Native American [S1, S2]. To make inferences about genomic

ancestry while utilizing the properties of genomic principal components (PCs), we used

SNPweight software with external genome-wide SNP reference panels in African (YRI),

European (CEU), East Asian (EA), and Native American (NA) [S3-S5]. The external

reference panels included data from HapMap 3 in addition to a collection of Native

American samples [S3, S4]. For inference, 364,452 autosomal SNPs were used. The

SNPweight software provided weights for these SNPs to project the genomic variation of

our sample onto the reference PCs (Figure S1). We then defined, for each participant, the

proportion of ancestry in the reference-inferred PCs space [S5], for each ancestral

population. The metrics were standardized with ranges from 0% to 100%. The centroid of

a specific reference group was set as 100% of the proportion of genetic ancestry in that
specific ancestral component. Increasing proportion of genetic ancestry in one component

represents decreasing genetic distance from that reference group (Figure S1).

Image acquisition, quality control, and processing

Neuroimaging data were acquired across study sites with 12 3-Tesla scanners. To ensure

consistency across scanners, a standardized multi-modality high-resolution structural

MRI protocol was implemented. Detailed protocols have been published elsewhere and

made available online [S1, S2, S6]. A standard PING imaging session included a 3D

T1-weighted, inversion-prepared, RF-spoiled gradient echo scan using prospective

motion correction, a 3D T2-weighted variable flip angle fast spin echo scan, a high

angular resolution diffusion imaging scan with integrated B0 distortion correction, and a

resting state blood oxygenation level-dependent fMRI scan. Acquisition protocols with

standardized pulse sequence parameters were installed on scanners at all study sites.

Individual sites uploaded DICOM images through the secure web-based application. The

uploaded data were automatically checked for completeness and protocol compliance.

Trained experts then reviewed each image for motion artifacts, excessive distortion,

operator error, or scanner malfunction. The quality ratings were input within 24 hours

from time of upload to allow re-scanning. Only images that passed quality control were

included in the current analysis.

T1-weighted imaging that was the main focus of the current analyses included a

sagittal 3D inversion recovery spoiled gradient echo T1-weighted volume. Distortions

caused by nonlinearity of the spatial encoding gradient fields were corrected with

predefined nonlinear transformations [S7]. Non-uniformity of signal intensity was

reduced with the nonparametric nonuniform intensity normalization method [S8]. To

facilitate standardized viewing and analysis of brain structure, images were rigidly

registered and resampled to alignment with an atlas brain with 1 mm isotropic voxels.

To reconstruct the cortical surface from T1-weighted images, a semi-automated

process was conducted using FreeSurfer software [S9, S10]. The automated process

corrects variation in image intensity due to RF coil sensitivity inhomogeneities and then

covers the cortical surface with a polygonal tessellation. Each participant’s cortical

surface was then mapped to a spherical atlas space based on the continuous cortical

folding patterns. The surface meshes for each individual were then re-tessellated to have

a uniform number of vertices per hemisphere (163,842 vertices as icosahedral order 7).

Afterward, the standardized atlas surface tessellation was transformed back into subject

space, deriving the 3-D Cartesian coordinates of each vertex on the cortical surface.

Through this process, the coordinates of cortical vertices can represent the variation of

cortical surface geometry for each individual while ensuring that each vertex represents

the same location relative to cortical folding patterns.

Demographic and covariates distribution

Table S1 shows the stratified distribution of participants whose data have passed quality

control. Because the genetic ancestry is a continuous scale instead of categorical, the

proportion showed in each cell is the weighted sum across the whole sample. As Table S1

shows, CEU represents a larger proportion of the participants. The distribution of the

proportion reflects the underlying population structure of the United States. The gender

distribution is also even across ancestries (Table S1, Figure S2). Age only explained less
than 1% of the variance of genetic ancestry, indicating no clear evidence for recruiting

bias in terms of age and ancestry. Household income also has limited correlations with

genetic ancestry, explaining less than 1% of the variance of genetic ancestry. To prevent

possible confounding effects, covariates listed here were included in all models.

Prediction models for genetic ancestry

We built models that use the coordinates of vertices of reconstructed cortical surfaces to

predict the proportion of genetic ancestry. We used three methods to prevent over-fitting

and derive unbiased estimates of model performance. First, we subsampled the cortical

surface tessellation from the original 163,842 vertices to 642 vertices per hemisphere.

There are three vectors for each vertex to represent the morphing directions in 3-D

Cartesian coordinates, yielding 3,852 variables in each model. Second, we used penalized

regression with L2 norms (ridge regression) to further reduce the problem of over-fitting.

Third, we used leave-one-out cross-validation to derive unbiased estimates of model

performance. Age, age squared, scanner, and gender were treated as covariates. Further

secondary analyses were carried out to examine the roles of surface area variability and

of simpler, global measures of brain morphology in the prediction of ancestry (see section:

Models Using Other Brain Imaging Measures).

Visualization of cortical surface morphing

For each of the 4 reference populations, proportion of genetic ancestry was regressed on

the 3-D directional vectors of each vertex. Age, age squared, gender, and scanner were

treated as nuisance covariates. The estimated linear coefficients provide the magnitude
and direction a vertex would move in 3-D space when the proportion of genetic ancestry

increases from zero to 100 percent. These linear predictions for the 3-D coordinates of

vertices on cortical surfaces were used to build the 3-D representation of the cortical

surface geometry. To visualize the morphological differences in each 3-D reconstructed

cortical surface, we built five predicted 3-D cortical surfaces, one for each ancestral

category and one as the ancestry-neutral cortical surface (25% proportion of ancestry in

all ancestral components). We then visualize the morphological differences between the

ancestry-neutral cortical surface and predicted 3-D cortical surface with 100% proportion

of genetic ancestry in each ancestral component. Surface morphing from ancestry-neutral

to the 100% ancestral surfaces was illustrated in both animations (Movies S1 to S4) and

in a color-coded still image (Figure 2). In the still image, the red-blue-green colors

represent the morphing directions in a standard 3-D space. Red represents variations in

the medial-lateral axis; green represents variations in the anterior-posterior axis; and blue

represents variations in the dorsal-ventral axis. The length of each line represents the

morphing magnitude of each vertex from ancestry-neutral cortical surface to the 100

proportion of genetic ancestry. In the animations, the smoothing was done on the surface

with a Gaussian kernel (SD: 0.6 mm). Smoothing reduced the jagged interpolation

between vertices, and the width of the smoothing kernel didn’t affect the folding patterns.

Models Using Other Brain Imaging Measures

Cortical surface geometry contains rich information about the shape of the human brain

that may be attributable to other related brain phenotypes; i.e., cortical surface area,

aspect ratios of brain, total brain volumes, and volumes of subcortical structures [S10,
S11]. We conducted a series of analyses to examine the predictive power of those

imaging measures. All models were evaluated with comparable parameter settings. The

results are summarized in Table 1. The aspect ratios included length, width, and height of

the cerebrum, width of the cranial base, width of the cranial vault, and their pairwise

ratios. The subcortical region-of-interest includes volumes of white matter, thalamus,

caudate, putamen, pallidum, hippocampus, amygdala, nucleus accumbens, and ventricles.

Genetic spatial model

To check if the proportion of genetic ancestry defined in the reference-inferred PC space

reflects spatial structure of genetic variation across the world, we did an additional

analysis with a parameterized genetic spatial model, using SPA software [S12]. The SPA

assumes that the SNP allele frequencies change gradually and proportional to the

geographical distance. Using whole genome SNP data, SPA can map each individual to a

geographical location based on their genotyping. We used the dataset from the Human

Genome Diversity Project [S13] to train a 3-D global model and then mapped our data

onto the 3-D globe based on our sample’s genomic information. Although our SPA

model uses independent reference populations and different ancestral categories, the

results are comparable with the PCA approach. Figure S1 demonstrates that the

proportion of genetic ancestry, to a certain degree, reflects the spatial structure of genetic

variation across the world.

Simulated neuroanatomical change along the Great Diaspora

We did an additional demonstration of the possible utility of modeling 3-D geometry

with genetic ancestry. We morphed the cortical surface geometry from Africa to Europe,

Europe to East Asia, and East Asia to America, following the known sequence of

dispersal of modern humans from Africa [S14]. The results are shown in Figure S3.
Supplemental References:

S1. Walhovd, K.B., Fjell, A.M., Brown, T.T., Kuperman, J.M., Chung, Y.H., Hagler,

D.J., Roddey, J.C., Erhart, M., McCabe, C., Akshoomoff, N., et al. (2012).

Long-term influence of normal variation in neonatal characteristics on human

brain development. Proc. Natl. Acad. Sci. U.S.A. 109, 20089-20094.

S2. Fjell, A.M., Walhovd, K.B., Brown, T.T., Kuperman, J.M., Chung, Y., Hagler,

D.J., Venkatraman, V., Roddey, J.C., Erhart, M., McCabe, C., et al. (2012).

Multimodal imaging of the self-regulating developing brain. Proc. Natl. Acad. Sci.

U.S.A. 109, 19620-19625.

S3. Reich, D., Patterson, N., Campbell, D., Tandon, A., Mazieres, S., Ray, N., Parra,

M.V., Rojas, W., Duque, C., Mesa, N., et al. (2012). Reconstructing Native

American population history. Nature 488, 370-374.

S4. The International HapMap 3 Consortium (2010). Integrating common and rare

genetic variation in diverse human populations. Nature 467, 52-58.

S5. Chen, C.Y., Pollack, S., Hunter, D.J., Hirschhorn, J.N., Kraft, P., and Price, A.L.

(2013). Improved ancestry inference using weights from external reference panels.

Bioinformatics 29, 1399-1406.

S6. The Pediatric Imaging, Neurocognition, and Genetics Study: Methods Used in

PING. (http://pingstudy.ucsd.edu).

S7. Jovicich, J., Czanner, S., Greve, D., Haley, E., van der Kouwe, A., Gollub, R.,

Kennedy, D., Schmitt, F., Brown, G., Macfall, J., et al. (2006). Reliability in

multi-site structural MRI studies: effects of gradient non-linearity correction on

phantom and human data. NeuroImage 30, 436-443.

S8. Sled, J.G., Zijdenbos, A.P., and Evans, A.C. (1998). A nonparametric method for

automatic correction of intensity nonuniformity in MRI data. IEEE transactions

on medical imaging 17, 87-97.

S9. Fischl, B., Sereno, M.I., and Dale, A.M. (1999). Cortical surface-based analysis.

II: Inflation, flattening, and a surface-based coordinate system. NeuroImage 9,

195-207.

S10. Dale, A.M., Fischl, B., and Sereno, M.I. (1999). Cortical surface-based analysis. I.

Segmentation and surface reconstruction. NeuroImage 9, 179-194.

S11. Chen, C.H., Gutierrez, E.D., Thompson, W., Panizzon, M.S., Jernigan, T.L., Eyler,

L.T., Fennema-Notestine, C., Jak, A.J., Neale, M.C., Franz, C.E., et al. (2012).

Hierarchical genetic organization of human cortical surface area. Science 335,

1634-1636.

S12. Yang, W.Y., Novembre, J., Eskin, E., and Halperin, E. (2012). A model-based

approach for analysis of spatial structure in genetic data. Nat. Genet. 44, 725-731.

S13. Cavalli-Sforza, L.L. (2005). The Human Genome Diversity Project: past, present

and future. Nature Reviews Genetics 6, 333-340.

S14. Li, J.Z., Absher, D.M., Tang, H., Southwick, A.M., Casto, A.M., Ramachandran,

S., Cann, H.M., Barsh, G.S., Feldman, M., Cavalli-Sforza, L.L., et al. (2008).

Worldwide human relationships inferred from genome-wide patterns of variation.

Science 319, 1100-1104.

S15. Cavallisforza, L.L., Menozzi, P., and Piazza, A. (1993). Demic Expansions and

Human-Evolution. Science 259, 639-646.