BWA parameters for mRNA-seq aligning against mRNA refseq

[kwicher]

Hello

I am sure similar questions were posted earlier but I could not find the definitive answer.

So the question is what parameeters would you use for mapping 1/16 chip data from single-cell mRNA-seq against mRNA refseq database. I am mostly interested in diferential expression of genes not SNPs, gene fusions, etc.

I used bwa with standard parameters "bwa aln -c -t 4 ..." and I have got extremely low mapping yield e.g. from total 40 291 303 reads, 8 529 466 reads were mapped to 6541 refseqs (of more than 36 000 refseqs in the data base).

Any idea how to obtain better mapping?

Thanks

K.

[colindaven]

What yields do you get mapping to the whole (human ?) genome, then counting reads after the mapping step ? The EdgeR package was quite good for this purpose for me and has a nice tutorial available online.

I think mapping rates of 50% are quite normal for Solid data, and in my human transcriptomes it seems about 30-50% may be mapping outside exons (Illumina data). I've not looked at this in detail so far though.