IUPAC aligners

[ohofmann]

Dear all,


been going through a fair amount of user guides and manuals, but so far I've only found one aligner able to handle IUPAC codes in reference sequences (MOSAIK, http://bioinformatics.bc.edu/marthlab/Mosaik).

Any pointers to additional tools which we could give a try would be very welcome!

Cheers, Oliver

[sparks]

Hi Oliver,

Novoalign handles IUPAC NA codes in the reference sequences. During index creation (it indexes the genome) it will create multiple k-mer index entries for k-mers with up to 2 ambiguous codes. This helps with the seeding part of the alignment.

The IUPAC codes are also used when constructing scoring matrix for the dynamic programming (Needleman-Wunsch) alignment. For instance a purine (R) aligns to an A or G (quality = 30) in the read with a penalty of 3 and to a C or T with a penalty of 30.
A B (C, T, or G) will align to C,G or T in read with P=0.3 and a penalty of 5 and to an A with penalty 30.
An N in the reference sequence will align to any base in the read with a score penalty of 6. This is why some Novoalign alignments extend into blocks of N's.

The above examples are based on read bases with Phred quality of 30. The actual penalties are adjusted according to the quality values.

Colin

[BaCh]

Any pointers to additional tools which we could give a try would be very welcome!

MIRA also handles IUPACs, be it for de-novo or mapping assemblies.

http://www.chevreux.org/projects_mira.html

B.

[ohofmann]

Thanks everyone for the feedback. Along with the regular aligners that's turning out to be quite a project for the comparison alone. Should be fun, and I'll report back on the results as soon as I can.

[lwhitmore]

Does any one know if CrossMatch the alignment tool with RepeatMasker can handle IUPAC codes