Multiple alignment: compare one sequence with many

Question

[mdihcroe]

Hi everyone! Does anyone know a tool to conduct multiple alignment so that there's one sequence (query) which is aligned to a list of sequences and in the results those letters are highlighted which are well-aligned with this query. I really like MUSCLE tool, but in this tool all sequences are equal and in the results it's hard to see how much all the sequences vary from the one I want them to be compared to.

UPD: it would also be very helpful if I could tell from the output how many bases of query sequence are matched to other sequences in alignment and choose best-aligned ones with a threshold e.g. at least 14 nucleotides from query sequence are same in the chosen best-aligned sequences

[GenoMax]

Try MEGA.

[natasha.sernova]

See this link:

http://www.mybiosoftware.com/tag/multiple-alignment

and the link below looks like your tool:

http://www.mybiosoftware.com/jphmm-compare-sequence-multiple-alignment-sequence-family.html

[Joe]

You can use any pairwise aligner of choice and concatenate the one-vs-many alignments to a new file. If you padded that file with gap characters, you could treat it like a normal multiple sequence alignment, and then use any consensus illustration tool you like as normal.

[hfahad82]

Use BioEdit for multiple sequance aligment

Answer 1

[Jean-Karim Heriche]

Use Clustal omega, it can align a sequence to a multiple sequence alignment (see the README file) or use the HMMER package: convert your multiple sequence alignment to HMM (which is what clustal does) with hmmbuild then use hhmalign.

[lieven.sterck]

I was actually thinking along the same line (with the HMMer approach) but how would you then highlight the well aligned ones?

[Jean-Karim Heriche]

I would use Jalview.

[mdihcroe]

I might misunderstand something but as I understand the pre-alignment of multiple sequences would represent kind of most similar patterns of the sequences. And when I align the query sequence to the pre-alignment, it would try to fit the sequence to these patterns. And ignore that maybe some of the sequences are more similar to the query sequence than to the others. But what I actually want to do is to align the query sequence to each sequence and then show visually which of the sequences aligns to the query better, has more common bases in same positions (that I imagine reminds of multiple alignment but I don't really want to know how similar the sequences among each other are). I'm sorry I might have explained my task vaguely in the first place. BLAST doesn't help since I've got very short sequences and BLAST ignores some matching positions, it shows only streched matching regions in the output.

[Jean-Karim Heriche]

It depends on how you define "align better". It seems that you want pairwise global alignments in which case BLAST is the wrong tool (BLAST means Basic Local Alignment Search Tool). Look for the Needleman-Wunsch algorithm (e.g. program needle in the EMBOSS package).