Update on 2013-11-26:
Archives bcgsc.ca_STAD.IlluminaGA_RNASeq.Level_3.1.3.0 and bcgsc.ca_STAD.IlluminaGA_RNASeq.Level_2.1.4.0 include updated analysis results for sample TCGA-BR-4294-01A-01R-1131-13 that are based on TCGA-BR-4294-01A-01R-1131-13_rnaseq.bam (cgHub upload date: 2013-11-26; analysis UUID: 081fa3ef-c3fd-43da-89b0-604ad36b46ee). The previous analysis results for said sample were based on a corrupt BAM and should not be used.

Update on 2014-02-28:
The BCGSC's v2 alignment pipeline now includes additional post processing that is designed to reduce the number of alignment artifacts. This post-processing identifies alignments with mismatches that were not reported in the SAMRecord's MD tag.  It corrects the SAMRecords CIGAR strings for reads that have minor alignment artifacts caused by incomplete soft clipping. For reads that have discrepancies in the MD tag simply due to the way bwa treats Ns in the reference, the post processing updates only the SAMRecord's MD string. For the small number of SAMRecords identified by the post processing pipeline as not falling into the first two cases, the records are simply converted to unaligned so that they no longer influence downstream analysis. The new pipeline reduces the number of false SNP calls, and has a minor impact on expression quantification.
