Aligning to a an ambiguous (IUPAC) reference in Mosaik

[lies]

Hi all,

We are currently trying to design a pipeline for viral sequencing with Illumina. For the alignment, we were considering using Mosaik with an ambiguous reference sequence. This sequence is constructed from 80 downloaded sequences. We also constructed an unambiguous sequence from the same data.

Funnily enough, aligning with respect to the unambiguous reference sequence gave us decent alignment stats (% of reads mapped were in the 80-98 range), whereas alignment percentages using an ambiguous reference sequence were in the 14-41% range.

We are currently using the following parameters: -hs 11 -mmp 0.08. This is all done in Mosaik v2.1.73.

Does anyone know whether Mosaik v2 still supports IUPAC? Should we be using some flag that I'm unaware of? Is the hashsize too big to deal with IUPAC coded references? Any pointers are welcome. It's not that we're not happy with the current results -- we just want to *know...

Thanks for your help,

Elise

[lwhitmore]

Elise,

Did you ever find an answer to this question I am currently using Mosaik 2.2.26 and not getting good results either.

Thanks for any help,
Leanne

[lies]

Hey Leanne,

Nope... I contacted the author and she said
"The first step of MOSAIK is to find hashes and the default parameters cannot handle hashes with IUPAC. It may be the reason why the aligned rate for IUPAC references is very low."

However, I never found out which non-default parameter should be turned on in order to use the IUPACs. You can try to contact them yourself?

We moved on to Bowtie2 and are using multiple rounds of alignment in which we alter the reference sequence.

Elise

[lwhitmore]

Elise,
Ok thanks for getting back to me. I contacted the author as well but I haven't heard back.

Thanks
Leanne