MSA following hmmalign/hmmsearch?

Question

[db534]

Hello,

My question is in regards to the hmmsearch/hmmalign commands of the HMMER program. As an output, both of these commands produce a sequence alignment with respect to the HMM that was used to query the sequence database. My question is, from the aligned domain sequences that can then be extracted, would it at all be necessary to use a conventional multiple-sequence algorithm to realign the domain sequences (extracted)?

My thinking is that by using the most accurate algorithms (e.g. MAFFT L-INS-i), a more reliable alignment would be produced -- in an analogous sense to still using an MSA algorithm to align the output of a BLAST search, even though the output is itself an alignment? Not sure whether this would be necessary or not.

Answer 1

[krinsman]

I don't know how to answer this question. I'm just wondering, what do you mean by "necessary"? I.e. do you mean that the MSA resulting from using HMMalign would be sufficiently inaccurate as to be useless? Or are you asking for specific contexts where the MSA produced from HMMalign would be expected to be more accurate than the MSA produced from a "de novo" MSA algorithm (like MAFFT-Linsi), or where the "de novo" MSA would be expected to be more accurate than the MSA produced by HMMAlign? Or something else?