problems with Bioscope output

[Robby]

Hello everyone,
I mapped my Solid data with Bioscope. In the csfasta files are around 142 million reads. This value is shown in the pairing statistics file as well. But in the bam file with the mapped and unmapped reads are together just 119 million reads. I checked that with the help of samtools. What happened to the other reads? Does Bioscope filter any reads? I couldn't find anything about that.

Furthermore the pairing statistics show more mapped pairs than I could find in the bam file. Could anybody explain that, please?

I would be very happy about some help or comments. Thanks

[KevinLam]

you do mean that you counted the mapped and unmapped files? and the numbers do not add up?
which version of Bioscope are you using? can you post the ini file?

[hanktu]

Hello Robby,

You mentioned that the input csfasta file contains roughly 142 million reads. As you know not all reads are mappable so you should expect some losses in the bam file. To be very specific, bam file will only have an entry for an unmapped read IF the sister read (the other half of the pair) is mapped. If both reads from a pair are not mappable, then the pair will not show up in the bam file. This should explain the discrepancy you are observing.

Regarding the differences between pairing statistics and bam file, could you post the stats from both approaches? It is difficult to troubleshoot without something more concrete.

Thanks

[Robby]

Thanks for your help and for parts of the solution. As hanktu mentioned all pairs of which both reads are mappable appear in the bam file with the mapped reads as two reads. If both reads from a pair are not mappable, then the 2 reads appear in the bam file with the unmapped reads. But if one read in a pair is mappable and the other not, then one entry appears in the bam file with the mappable reads and no entry in the file with the unmappable reads. So the unmapped read of the pair doesn't appear anywhere. The problem was, that Bioscope considers pairs and I counted reads.

But nevertheless 80-160 reads (or 40-80 pairs) are missing. In the Bioscope statistics they are counted as mapped. Do you have any ideas for that as well?

[hanktu]

What version of BioScope?

Hi Robby,

What version of BioScope are you running? Can you copy and paste some lines from the pairing stats file as well as the numbers you get from the bam file?

Thanks

[Fabian]

Could this be caused by redundantly mapped reads? If you're creating the bam file from 'primary' as the config setting (I think this is the default), then there will be reads that map to many locations that are 'mappable' but will not appear in your bam files unless they can be rescued by a primary/uniquely mapping mate.