[nilshomer]

when i use samtools.pl varFilter to filter the variant and snps, it seemed the program have a cap on the maximum read depth, it filtered most of the candidates that has higher coverage that should not be filtered, however when I try to redefine the D through command line arg, the program return nothing after filtering, any one know how to fix it?

Thanks

Try posting to the samtools help list ([email protected]) or creating a new thread, since this is not related to the original question.

[jsun529]

Thank you very much for all your help, since I am very new to this BFAST program, The program runs pretty fast and with the new version it runs pretty smoothly. For my test run on solid, when I use the sam tools to call variants, the result was not what I expected, I think it comes to both side, one is the samtools, another is I think more critical as how to create proper indexes, for snp and indels? Looks like with new version this is the only parameter left to the user? Thanks

[nilshomer]

Thank you very much for all your help, since I am very new to this BFAST program, The program runs pretty fast and with the new version it runs pretty smoothly. For my test run on solid, when I use the sam tools to call variants, the result was not what I expected, I think it comes to both side, one is the samtools, another is I think more critical as how to create proper indexes, for snp and indels? Looks like with new version this is the only parameter left to the user? Thanks

The same parameters are available, they just now default to the recommended settings for whole-genome human resequencing. The indexes are important, and depend on read length, error rates, and polymorphism rate (snp and indel). What read length, error rate, platform, and polymorphism rate are you considering?

[colindaven]

Hi Nils,

I found Bfast challenging to get into as well. I saw that as someone who has been using Maq, Mosaik, Bowtie and so on for quite some time now.

There is plenty of detail in the manual but the program expects you to set parameters right from the off, without any suggestion of defaults. Now it may well be this is complicated and reference dependent, but to encourage users I would strongly recommend

a) a short sample reference, say a genomic island or 500k of a genome
b) a few reads which can be mapped onto this.

There is plenty of public data about now.

Then I would have a group of command lines in place of the "quick tutorial" links.

For users unfamiliar with installing from c source instructions could be provided as well.

I m sure potential users would appreciate the effort involved in doing this

Colin

[nilshomer]

Hi Nils,

I found Bfast challenging to get into as well. I saw that as someone who has been using Maq, Mosaik, Bowtie and so on for quite some time now.

There is plenty of detail in the manual but the program expects you to set parameters right from the off, without any suggestion of defaults. Now it may well be this is complicated and reference dependent, but to encourage users I would strongly recommend

a) a short sample reference, say a genomic island or 500k of a genome
b) a few reads which can be mapped onto this.

There is plenty of public data about now.

Then I would have a group of command lines in place of the "quick tutorial" links.

For users unfamiliar with installing from c source instructions could be provided as well.

I m sure potential users would appreciate the effort involved in doing this

Colin

Great suggestions! What was the last version you used? I say that because an overhaul was made between 0.5.x and 0.6.x to improve the user interface, manual, and set default/recommended parameters.

There is a quick tutorial at the end of the manual giving examples and the recommended parameters (the program now has defaults). Does this address the issue?