Alignment of sequences composed by units of multiple characters

Question

[benjap700]

Dear all

I'm trying to aling short sequences composed by units of 3 characters. I use 3 characters because I need to codify at least 6000 different entities (enzymes in this case). An example of this kind of sequences could be: ABC ADC ADX ... ( I put an space after each 3 letters, in fact this sequence is ABCADCADX).

Probably exist any software that allow to choose the number of characters that in fact be an unique unit of alignment. Also I need to introduce a my own similarity matrix (different that the usual like Blosum62)

Please, do you know any software that could help?

Thank you in advance

Benjamí

[Joe]

Am I understanding this right that you're trying to do codon-aware alignment?

If not, you might need to explain further, perhaps showing some ideal output.

[benjap700]

Thank you for your quick answer, Joe

No, I'm not trying to align codons, but the problem is similar: I need to codify my units of alignment with 3 characters, because I have to codify 6000 entities (enzymes in this case). Then, my sequences are composed by units of 3 letters and the software must consider this fact.

Answer 1

[Asaf]

You can use EMBOSS needle (for global) or water (local) alignment. They allow a custom substitution matrix as input using the -datafile option. You can use the command:

embossdata -fetch -file EDNAFULL

to copy the DNA default matrix to your working directory and change it to use only three letters by setting the scores to a very low number. See more here: http://emboss.sourceforge.net/apps/release/6.6/emboss/apps/water.html

[benjap700]

Thank you, Asaf All the best :)