This FGWAS package is developed by Chao Huang and Hongtu Zhu from the BIG-S2 lab.

FGWAS (Functional Genome Wide Association analysiS) is a Python coding based package for imaging genetic analysis. Functional phenotypes (e.g., subcortical surface representation), which commonly arise in imaging genetic studies, have been used to detect putative genes for complexly inherited neuropsychiatric and neurodegenerative disorders. However, existing statistical methods largely ignore the functional features (e.g., functional smoothness and correlation). The aim of this toolbox is to develop a functional genome-wide association analysis framework to efficiently carry out whole-genome analyses of functional phenotypes. FGWAS consists of three components: a multivariate varying coefficient model, a global sure independence screening procedure, and a test procedure. Compared with the standard multivariate regression model, the multivariate varying coefficient model explicitly models the functional features of functional phenotypes through the integration of smooth coefficient functions and functional principal component analysis. Statistically, compared with existing methods for genomewide association studies (GWAS), FGWAS can substantially boost the detection power for discovering important genetic variants influencing brain structure and function. Simulation studies show that FGWAS outperforms existing GWAS methods for searching sparse signals in an extremely large search space, while controlling for the family-wise error rate.

Command Script

A.Pipeline: run the following scripts in order in python3: 0.preparation 0.1 modify PATH in the following 3 scripts to your path that contains all python scripts in this package 0.2 create the file directory structure (as in section B) 0.3 prepare input data (as under section B:data) 1_gsis.py [GSIS] output:

• GSIS_all.txt [each SNP per row for all SNPs, with columns: 1.SNP chromosome, 2.SNP position (bp), 3.global test statistic, and 4.p-value]
• GSIS_top.txt [each SNP per row for top SNPs (sorted from most significant to less significant), with columns: 1. SNP name, 2.SNP chromosome, 3.SNP position (bp), 4.-log10(global p-value)] 2_bstp.py [Bootstrap: can submit multiple jobs, each running this script for b bootstrap samples] output: max_lstat_bstp_i [bootstrap local test statistics] max_gstat_bstp_i [bootstrap global test statistics] max_area_bstp_i [bootstrap cluster-size statistics] combine output: cat max_lstat_bstp_* > max_lstat_bstp.txt cat max_gstat_bstp_* > max_gstat_bstp.txt cat max_area_bstp_* > max_area_bstp.txt 3_test.py [Local Test & Cluster-Size Analysis] output: local_pv_raw.txt [raw p-values at each voxel as local tests] local_stat.txt [local statistics at each voxel] local_pv_adj.txt [p-values at each voxel adjusted for # SNPs and # voxels] cluster_pv.txt [p-values for cluster-size analysis for each SNP] 4.[optional] global test: calculate p-values adjusted for the number of SNPs (Ng) based on column 3 of GSIS_all.txt and its distribution max_gstat_bstp.txt

B.Directory structure under PATH:

• data [input data; all w.o. row names nor column names] 1.img_data_left.mat, img_data_right.mat [m x n x l matrix of image response; m: # responses, n: # sumjects, l: # voxels] 2.coord_data_left.txt, coord_data_right.txt [l x d matrix of image coordinates; d: dimension of image (e.g. 2 for 2D image, 3 for 3D image)] 3.design_data.txt [n x p matrix of design matrix (w.o. 1st column of 1's; the 1st column of 1's will be added in code)] 4.var_type.txt [vector of length p; the ith element is 1 if the ith column (covariate) in design_data.txt is numeric, and 0 otherwise] 5.snp_data.txt [n x Ng matrix of SNP dosage levels] 6.snp_info.map [Ng x 3 matrix of SNP info, w. each row per SNP; with columns: 1. SNP chromosome, 2.SNP name, 3.SNP position (bp)] 7.img_size.txt [length-d vector, indicating the size of the image (e.g. '40 50 30' for a 40x50x30 3D image)] 8.img_idx.txt [length-l vector, containing the order of the voxels (e.g. 1:l)]
• res [contains final results from 1_gsis.py and 3_test.py]
  • left [assume there are left side and right side such as for hippocampus. Otherwise, modify the file structure accordingly.]
  • right [similar to 'left'] ...
• vars/ [contains intermediate outputs]
  • left/
  • right/
• bstp [contains boostrap statistics (output from 2_bstp.py)]
  • left/
  • right/
• work [optional directory to hold your job-running log files]