quantitative measure of coverage and reference genome issue

[sara_]

Hi all,

I need to analyze paired ends reads for a target re-sequencing experiment in which the repeated regions of my targets were masked.

I have performed the alignment with BWA and converted the export of BWA alignment in a SAM file. Then using SamTool I have removed duplicated reads and run the SNP and indel calling. Does anyone know how I can get a quantitative measure of the coverage of my target regions?

Moreover, I have few theoretical questions regarding the reference genome that should be used. Even if it is a target re-sequencing, is it correct to align my reads against the full genome? Indeed I have aligned against the full genome, but it also contained the clone contigs that can't be confidently placed on a specific chromosome (e.x. chrUn_gl000225 chrUn_gl000222, chr6_qbl_hap6, chrUn_gl000225). In your opinion is it correct or is it too much a conservative approach? Finally, as I have performed a target sequencing in which I masked the repeated regions, is it correct to align my reads directly against the masked genome?

Many thanks to anyone could help,

Sara

[biznatch]

I think it's better to align to an unmasked genome and use a setting that will do something like, if a sequence aligns to more than once place, discard it. If you align to a masked genome then sequences which should actually align to a repetitive region may get "forced" to align somewhere else, depending on your alignment settings. Also, from what I read somewhere on here that the masked genome isn't 100% correct and may mask some non-repetitive sequences.

Regarding the unaligned contigs, when I align to the mouse genome I include them because there are some genes on them I'm interested in. I'd think it's best to include them when aligning, but I haven't done any alignments to the human genome so I can't say for sure.