Copy Number Variants from SSC Collection ~ 2500 families by two Methods (XHMM and Conifer)

XHMM was run on a set of realigned BAM files from the SSC collection (see NDAR Study 334 for BAM files) using the attached scripts. These scripts calculate depth of coverage using GATK, pull the GATK output from an instance on NDAR's cloud, merge the output of GATK into a single matrix, process the read depth matrix (filter, center), normalize the matrix using principal component analysis (PCA), process the normalized read depth matrix (filter, z-score), run a hidden markov model (HMM) on this matrix to identify CNVs in the normalized data, and generate family level vcfs from the xhmm data. XHMM produces as output coverage summary tables produced by GATK (sample_interval_statistics, sample_interval_summary, sample_summary, sample_statistics), principal component data files, a genotyped CNV output VCF file, and some example plots and graphics. For this study, the GATK output is available.

Additional information about XHMM is available here: http://atgu.mgh.harvard.edu/xhmm/tutorial.shtml