Determine Coverage of SNV Loci

[Hecate08]

Hello,

We use instead of the normal workflow from snpir STAR pass-2 and the GATK best practises to create a bamfile. You recommend to use bwa-mem to align to the normal human reference genome to get undeflated coverage values by using genomecov -bg. Is it also possible to use genomecov -bg -split on our bamfile so we don't have to realign it? Will this make a big difference?

Thank you
Hecate

[biomrpaul]

Hi Hecate,

How will the the -split functionality allow you to rescue read count information that has aligned to the SNPiR splice junction reference? If it is able to do so, that seems like a great alternative to realigning the sample.

Let me know,
Thanks,
Matt Paul

[Hecate08]

Hello Matt Paul,

This will not help with the SNPiR splice junction reference since they use a combination of the normal human reference and "exonic sequences surrounding all currently known splice junctions". But the Star alignment only uses the normal human reference genome and an uncompressed suffix arrays index. So the read count will not be split between the reference and the exonic sequences surrounding splice junctions as it happens in the SNPiR alignment. I hope I understand this in the correct way.

Thanks,
Hecate

[biomrpaul]

Hey Hecate,

The authors of the SNPiR paper recommend their proposed pipeline for SNV calling from RNA-seq data. This includes aligning to the pseudo-reference of known splice junctions. If you are using a different pipeline for variant calling and will not being using this step, calculating genomic coverage directly from your alignment should be fine.

Cheers,
Matt