[JohnK]

For the BFAST+BWA mode, you align the 50bp end with BFAST and 35bp end with the BWA (within BFAST+BWA). This substitutes the "match" step. You then input the two BMF files into the "localalign" step, and proceed from there as before. There is not much documentation on this, and I would be happy to incorporate your feedback into a wiki page on the BFAST site going forward. There is always the option to post on [email protected], as there are many users that can help out there too.

For SNP calling, the implementation in SAMtools is sufficient, but is not color-aware like dibayes, though I have not seen enough evidence to convince me which is better. The mapper used by Bioscope is much improved, as such would be comparable to bfast. Make sure that you do appropriate filtering after SNP calling. A good guide to follow is what the 1000 Genomes project does. Please post your results, as they would benefit the community. I always appreciate your posts.

Hi, Nils.

I was perusing their website and wasn't sure if I was at the right place, but it was the analysis page. Looking there, I understand they use multiple SNP callers to get the overlap and this reduces up to 30% of false positives for low with coverage? Would you say in your own opinion that it's vital to have two different assemblies from different programs when comparing the SNP calls or would two separate SNP callers on one alignment dataset be, well, good? It's necessary I make these calls in the quickest amount of time.

I need to align my reads in a much quicker method as BioScope has a very poor runtime, but I do appreciate BioScope's high fidelity in percentage of mapped reads. What additional percentage of mapping would you expect with BFAST? I think outside of BioScope, your's is the best open-source alternative. As far as runs are concerned, I find it very difficult to gauge a run outside of error rate. Do you have any recommendations for gauging run quality other than heat maps and cycle scans? It seems there are so many variable factors that go into the prep that two runs may appear good quality and yet it doesn't map over > 80 % with the same parameters as a similar run. I'd love to hear your thoughts on all these topics, as you have much experience? Thank you.

[nilshomer]

Given your limited time, do what you think is best. Other SOLiD aligners include BWA and SHRiMP2. The latter is much faster than the previous version and is worth considering.

[qtnguyen]

Filtering after BFAST and SNP calling

> Make sure that you do appropriate filtering after SNP calling. A good guide to follow is what the 1000 Genomes project does. [/QUOTE]

Could yo please elaborate, what would be an appropriate filtering procedure after SNP calling? How one can perform this?

Thanks,