Neuroimage. 2010 Nov 15;53(3):1160-74. doi: 10.1016/j.neuroimage.2010.02.032. Epub 2010 Feb 17.
Voxelwise genome-wide association study (vGWAS).
Stein JL,
Hua X,
Lee S,
Ho AJ,
Leow AD,
Toga AW,
Saykin AJ,
Shen L,
Foroud T,
Pankratz N,
Huentelman MJ,
Craig DW,
Gerber JD,
Allen AN,
Corneveaux JJ,
Dechairo BM,
Potkin SG,
Weiner MW,
Thompson P;
Alzheimer's Disease Neuroimaging Initiative.
Weiner M, Aisen P, Weiner M, Aisen P, Petersen R, Jack CR Jr, Jagust W, Trojanowki J, Toga AW, Beckett L, Green RC, Gamst A, Saykin AJ, Morris J, Potter WZ, Green RC, Montine T, Petersen R, Aisen P, Gamst A, Thomas RG, Donohue M, Walter S, Jack CR Jr, Dale A, Bernstein M, Felmlee J, Fox N, Thompson P, Schuff N, Alexander G, Jagust W, Bandy D, Koeppe RA, Foster N, Reiman EM, Chen K, Mathis C, Morris J, Cairns NJ, Taylor-Reinwald L, Trojanowki J, Shaw L, Lee VM, Korecka M, Toga AW, Crawford K, Neu S, Beckett L, Harvey D, Gamst A, Kornak J, Saykin AJ, Foroud TM, Potkin S, Shen L, Kachaturian Z, Frank R, Snyder PJ, Molchan S, Kaye J, Dolen S, Quinn J, Schneider L, Pawluczyk S, Spann BM, Brewer J, Vanderswag H, Heidebrink JL, Lord JL, Petersen R, Johnson K, Doody RS, Villanueva-Meyer J, Chowdhury M, Stern Y, Honig LS, Bell KL, Morris JC, Mintun MA, Schneider S, Marson D, Griffith R, Clark D, Grossman H, Tang C, Marzloff G, deToledo-Morrell L, Shah RC, Duara R, Varon D, Roberts P, Albert MS, Kozauer N, Zerrate M, Rusinek H, de Leon MJ, De Santi SM, Doraiswamy PM, Petrella JR, Aiello M, Arnold S, Karlawish JH, Wolk D, Smith CD, Given CA 2nd, Hardy P, Lopez OL, Oakley M, Simpson DM, Ismail MS, Brand C, Richard J, Mulnard RA, Thai G, Mc-Adams-Ortiz C, Diaz-Arrastia R, Martin-Cook K, DeVous M, Levey AI, Lah JJ, Cellar JS, Burns JM, Anderson HS, Laubinger MM, Apostolova L, Silverman DH, Lu PH, Graff-Radford NR, Parfitt F, Johnson H, Farlow M, Herring S, Hake AM, van Dyck CH, MacAvoy MG, Benincasa AL, Chertkow H, Bergman H, Hosein C, Black S, Stefanovic B, Caldwell C, Hsiung GY, Feldman H, Assaly M, Kertesz A, Rogers J, Trost D, Bernick C, Munic D, Wu CK, Johnson N, Mesulam M, Sadowsky C, Martinez W, Villena T, Turner RS, Johnson K, Reynolds B, Sperling RA, Rentz DM, Johnson KA, Rosen A, Tinklenberg J, Ashford W, Sabbagh M, Connor D, Jacobson S, Killiany R, Norbash A, Nair A, Obisesan TO, Jayam-Trouth A, Wang P, Lerner A, Hudson L, Ogrocki P, DeCarli C, Fletcher E, Carmichael O, Kittur S, Borrie M, Lee TY, Bartha R, Johnson S, Asthana S, Carlsson CM, Potkin SG, Preda A, Nguyen D, Tariot P, Fleisher A, Reeder S, Bates V, Capote H, Rainka M, Hendin BA, Scharre DW, Kataki M, Zimmerman EA, Celmins D, Brown AD, Pearlson G, Blank K, Anderson K, Saykin AJ, Santulli RB, Englert J, Williamson JD, Sink KM, Watkins F, Ott BR, Stopa E, Tremont G, Salloway S, Malloy P, Correia S, Rosen HJ, Miller BL, Mintzer J, Longmire CF, Spicer K.
Source
Laboratory of Neuro Imaging, Department of Neurology, University of California, Los Angeles School of Medicine, Neuroscience Research Building 225E, 635 Charles Young Drive, Los Angeles, CA 90095-1769, USA.
Abstract
The structure of the human brain is highly heritable, and is thought to be influenced by many common genetic variants, many of which are currently unknown. Recent advances in neuroimaging and genetics have allowed collection of both highly detailed structural brain scans and genome-wide genotype information. This wealth of information presents a new opportunity to find the genes influencing brain structure. Here we explore the relation between 448,293 single nucleotide polymorphisms in each of 31,622 voxels of the entire brain across 740 elderly subjects (mean age+/-s.d.: 75.52+/-6.82 years; 438 male) including subjects with Alzheimer's disease, Mild Cognitive Impairment, and healthy elderly controls from the Alzheimer's Disease Neuroimaging Initiative (ADNI). We used tensor-based morphometry to measure individual differences in brain structure at the voxel level relative to a study-specific template based on healthy elderly subjects. We then conducted a genome-wide association at each voxel to identify genetic variants of interest. By studying only the most associated variant at each voxel, we developed a novel method to address the multiple comparisons problem and computational burden associated with the unprecedented amount of data. No variant survived the strict significance criterion, but several genes worthy of further exploration were identified, including CSMD2 and CADPS2. These genes have high relevance to brain structure. This is the first voxelwise genome wide association study to our knowledge, and offers a novel method to discover genetic influences on brain structure.
Copyright 2010 Elsevier Inc. All rights reserved.
- PMID:
- 20171287
- [PubMed - indexed for MEDLINE]
- PMCID:
- PMC2900429
Free PMC ArticleFig. 1
The theoretical and observed distributions of the minimum P-value across voxels. (a) The normalized histogram of the observed minimum P-values is shown. Lines represent the PDF of the Beta(1, 275575) distribution (solid line) based on Meff and the Beta(1, 448293) distribution (dashed line) based on the number of measured markers. (b) The Q–Q plot shows the observed P-values plotted against those expected from the Beta(1, 275575) (blue dots). The black line gives a purely null distribution. The observed data matches well with that expected by the Meff based null distribution. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 2
A histogram and quantile–quantile plot for the “corrected” P-values (Pc-values). (a) The histogram shows the Pc-values approximately follow a uniform distribution. (b) The Q–Q plot shows the expected ordered −log10(Pc-values) as drawn from a uniform distribution plotted against the observed ordered −log10(Pc-values) as blue dots. The black line shows the null distribution. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 3
The cumulative distribution function of corrected P-values. The cumulative distribution function of Pc-values is shown (red) with two lines representing thresholds of q=0.50 (blue), and q=0.05 (green). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 4
The significance of the most associated SNP at each voxel. Each image represents slices through the brain at 8 mm intervals from inferior to superior. The top of the page represents anterior of the brain and the bottom of represents posterior. The images are in radiological convention (left of the image is the right side of the subject). Each voxel is colored by the –log10 of the P-value of the genetic association at each point (warmer colors are more strongly associated). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 5
The significance of the most strongly associated SNP at each voxel in a single permuted dataset. Each image represents slices through the brain at 8 mm intervals from inferior to superior. The top of the page represents anterior of the brain and the bottom of represents posterior. The images are in radiological convention (left of the image is the right side of the subject). Each voxel is colored by the –log10 of the P-value of the genetic association at each point (warmer colors are more strongly associated). The same color scale is used from Fig. 4 for comparisons. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 6
The locations of association for the 5 most associated SNPs. Slices through the MDT are shown in regions where the indicated SNP is the most associated at the voxel (red). The SNPs have effects on brain structure beyond the red colored voxels, but these voxels are associated with the labeled SNP more than any other. The slices through the MDT are every 4 mm and go from inferior (left of page) to superior (right of page). The images are in radiological convention (left of the image is the right side of the subject). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Fig. 7
The minimum number of subjects needed to replicate the findings for the top 5 most associated SNPs was estimated with a resampling approach. Subjects were randomly removed from each of the diagnostic categories until none was left in a category, and the association P-value of the SNP was calculated. This process was repeated 1000 times, to estimate 95% confidence intervals (red lines). The median P-value of the repetitions for each number of subjects removed is shown as the solid black line. The blue line shows the replication threshold for the first 5 SNPs, a Bonferroni corrected P-value of 0.01. The dotted blue line shows the estimated minimum sample size that would be required to detect a replication of the finding with 95% confidence (N=312 for rs2132683; N=263 for rs713155; N=291 for rs476463; N=299 for rs2429582; N=319 for rs9990343). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Neuroimage. 2010 November 15;53(3):1160-1174.
Publication Types
MeSH Terms
Grant Support
Full Text Sources
Other Literature Sources
Research Materials