The average sequencing depth was 50 million paired reads per sample. All reads were originally aligned by Tophat31 to the whole human genome reference (hg19) with Bowtie1 as the aligner. Several Picard metrics (http://broadinstitute.github.io/picard/) were collected from alignment results. Based on those Picard metrics, we implemented a paralleled and automatic RNAseq pipeline, in order to achieve higher quality of alignment and better estimation on gene expression levels. This pipeline includes identifying and trimming low quality bases (Q10) from beginning and end of each reads, identifying and trimming adapter sequencing from reads, detecting and removing rRNA reads and aligning reads to a transcriptome reference by a non-gap aligner (Bowtie1). The expression levels of gene and transcripts were estimated by RSEM package32. The Gencode V14 annotation were used by RSEM in the quantification process. Fragments Per Kilobase of transcript per Million mapped reads (FPKM) values were the final output of our RNA-Seq pipeline. 638 subjects passed QC from these two batches of samples.
