cellranger虽然是10x官方软件也未必得全信它
最近在整理各个癌症的单细胞转录组数据，发现 Cell. 2018 Aug 23;题目是：Single-Cell Map of Diverse Immune Phenotypes in the Breast Tumor Microenvironment. 文章的数据量本身就很大：
We profiled 45,000 immune cells from eight breast carcinomas, as well as matched normal breast tissue, blood, and lymph nodes, using single-cell RNA-seq.
Analysis of paired single-cell RNA and T cell receptor (TCR) sequencing data from 27,000 additional T cells revealed the combinatorial impact of TCR utilization on phenotypic diversity.
里面清清楚楚写到作者认为cellranger虽然是10x官方软件也未必得全信它：
We note that Cell Ranger, the most commonly used 10× pipeline, carries out an extreme version of this redesign: it removes any gene that is not protein coding. We believe that this is too harsh: it excludes numerous transcribed pseudogenes and lincRNA which have been previously shown to be expressed, have biological functionality, and be detectable in scRNA-seq.
如下图：
也就是说，作者认为，这个10X仪器的单细胞转录组数据走cellranger流程，其实是有一点问题的。
如果你关心的是走cellranger流程本身，我们在单细胞天地多次分享过流程笔记：
拿到表达矩阵后再走Seurat流程哦，两个流程基本上解决了10X仪器的单细胞转录组数据分析的80%基础内容。
作者的解决方案
首先自定义STAR比对软件参数
作者自己使用了STAR比对10X仪器的单细胞转录组数据，其实就是走cellranger流程，因为cellranger里面封装的就是STAR比对软件，但是作者参数选择是：Alignment parameters used are as follows: —outFilterType BySJout, —outFilterMultimapNmax 100, —limitOutSJcollapsed 2000000 —alignSJDBoverhangMin 8, —outFilterMismatchNoverLmax 0.04, – alignIntronMin 20, —alignIntronMax 1000000, —readFilesIn fastqrecords, —outSAMprimaryFlag AllBestScore, —outSAMtype BAM Unsorted
也就是说，作者允许每个测序的reads可以有多达20个位置的比对情况输出，而且没有成功比对的reads也会输出一条记录。
然后修复那些多比对情况的reads
既然作者允许每个测序的reads可以有多达20个位置的比对情况输出，那么就得给这些测序的reads一个归属。因为10X仪器的单细胞转录组数据本来就只有3端，而3端的序列本来就是比参考基因组其它区域更加同源。作者考虑了Kallisto (Bray et al., 2016) and EM approaches, such as RSEM (Li and Dewey, 2011), 两个方法，但是后面自己开发的方法有点复杂：