Alignment reads

[pingu]

Hello!
Do you suggest to align reads against whole genome or only chromosome?
I remembered that I have read that it is better to use whole genome (but I do not find the paper, if there is a paper when I can find it, could you send me the link?).
Thank you very much in advance

Best regards

[GenoMax]

Are you thinking of aligning to "known" transcriptome vs whole genome for RNAseq data?

Aligning to individual chromosomes (and combining that data) is no different would be roughly equivalent vs aligning to whole genome in one go.

[SylvainL]

Aligning to individual chromosomes (and combining that data) is no different vs aligning to whole genome in one go.

It depends if you remove multiple reads...

[GenoMax]

I was thinking of a simple case based on @pingu's description (against whole genome or only chromosome). Total data there is the same.

But as you point out, the alignments would be (roughly) equivalent only with specified caveats e.g. keeping all alignments when reads multimap irrespective of the number of alignments.

[pingu]

Thank very much for your response. I am working with DNA, in particular with amplicons for BRCA1 and BRCA2, and I have the reference sequence (brca1 and brca2) but I think that it is better to align with whole genome in order to avoid to get false alignments. Is it true?
Thank you for you help!

[GenoMax]

What downstream analysis are you planning to do?

[pingu]

I would like to find variants, in a diagnostic sector (I am working in a hospital).

[Brian Bushnell]

For whole-genome shotgun it's important to align to the whole genome, or you will get false-positive hits. For amplicon sequencing, though, I doubt it would make much difference. I suggest you map a few to the whole genome to see if they hit anything other than the intended target (due to pseudogenes or homologous sequences that may also be caught by the same primers). If all the reads only map on-target, then there's no point in mapping against the whole genome.