Cell-specific gene-expression profiles and cortical thickness in the human brain

Neurobiological underpinnings of variations in cortical structure, such as cortical thickness, in the human brain are largely unknown. In this data report, we describe a method to evaluate the contribution of nine neural cell-types in explaining inter-regional variations in cortical structure measured with FreeSurfer in participants from a study cohort, by correlating the cortical-structure data and gene-expression data from the Allen Human Brain Atlas, summarized into the FreeSurfer cortical regions and validated using the BrainSpan Atlas. The resulting graphical display and re-sampling based association test results can give new insights into the underlying neurobiological mechanisms of cortical structure variation.<br><br>Files available: Detailed description of each file is included in ‘README.doc’<br><br>AllenHBA_DK_ExpressionMatrix.tsv - Tab separated file containing correlation to the median values across the donors (left hemisphere only, column named ”Average donor correlation to median”) and gene expression values across the 68 FreeSurfer cortical regions (columns)<br><br>DKRegionStatistics.tsv - Tab separated file characterizing the FreeSurfer cortical regions. This file lists how many donors contribute to each region, Allen Brain Atlas samples per region and alternative identifiers.<br><br>BrainSpanToHBARegionMapping.csv - Comma separated file characterizing the 11 cortical regions that are commonly available for the Allen Brain and Brainspan Atlases.<br><br>Reference_Consistent_Genes_ObtainedBy2StageFiltering.tsv - Tab separated file containing correlation coefficients ("correlation"), the corresponding 1-sided p-value ("pvalue") for gene-expression levels between Allen and Brainspan Atlases for the 11 cortical regions that are available for both atlases, and cell-type ("CellType") assignments based on Zeisel et al. (2015) for the 2,511 consistent genes obtained by applying the following 2-stage procedure.<br><br>Profile_males_corcoef_thickness_age.csv - Comma separated file containing an example phenotype-profile (users need to format, like this example, and provide their own cohort-specific file).<br><br>GetBrainSpanCorrelations.R - R script to calculate and test correlation for gene expression profiles from Allen Human Brain vs. Brainspan Atlases.<br><br>LoadBrainSpanExpressionFunctions.R - R script defining a function to load expression data ‘Exon microarray summarized to genes’ available from Brainspan Atlas download website (http://brainspan.org/static/download.html).<br><br>GetGOGroupsFilteredHumanLists.R - R script to run GO analysis, where the foreground gene set is each Zeisel cell type panel is used as a foreground gene set, and the background set is the reference set with the 2,511 genes that pass the 2-stage filtering procedure.<br><br>GetExpressionPhenotypeCorrelations.R - R script to calculate expression-phenotype correlations and produce a 3-row-by-3-column graphical display of cell-type-specific empirical distributions of the resulting correlation coefficients with corresponding (1- significance level')×100% critical values, which are unadjusted for multiple testing.<br>