finding novel sequences : blastn or alignment against nr database ?

Question

[Farbod]

My Dear Expert Friends, Hi. ( I'm not native in English so, be ready for some possible language flaws).

I have read this topic, and it seems that "The Idea of that is to find novel sequences".

my question is this that which one is a better strategy for non-model animals RNA-seq to find novel sequences:

1- Running local blastn of transcriptome de-novo assembly against nt database

2- or aligning millions of reads against the NCBI nt database using BBMap

Please offer a short description that why you have chosen your preferred choice.

Thank you in advance.

Answer 1

[Brian Bushnell]

Blastn is not splice-aware, so it won't give full-length alignments. You'd have to spend some effort parsing and stitching together the split local alignments for each transcript to find out what portion of it was not covered. So, it might make more sense to translate the transcriptome to amino acids and blast against nr, so you don't have to deal with splicing.

I think, overall, the best strategy (in terms of sensitivity and resource requirements) might be to blast the assembled transcriptome in amino-acid-space, and use BBMap for just those reads that do not map to the transcriptome.

[Farbod]

Dear Brian Bushnell , Hi and thank you.

about the " use BBMap for just those reads that do not map to the transcriptome"

Do you mean to map them to nt database or mapping to some close species reference genome ?

[Brian Bushnell]

I meant, map them to nt... but it really depends on what you're trying to accomplish, and whether you have reference genomes of related species. If the goal is to see if you have sequences that do not appear in nt, then mapping to nt would be the way to go.