bbsplit not using all reads in library

Question

[ghd1998]

I have RNA-seq files which I am wanting to split based on mapping to reference sequences. I am using bbsplit to map to the sequences and output separate mapping files however I noticed that not all reads in my files are mapped using this method. My read file has 9654349 reads but each time bbsplit only uses 6233783 reads - is there a way for me to force all reads to be mapped?

When I use kmer splitting in bbduk to map to only one my reference sequences all of the reads are used so I am wondering if there is a flag or something I am missing which will allow me to split based on multiple reference sequences at once.

Thanks for your help in advance!

[lieven.sterck]

without knowing the details I can suggest to have a look at the parameters controlling how to deal with ambiguous/multi mappings. Perhaps that can explain the missing reads? Not sure what the default is but it can be set to ignoring ambiguous mappings which could result in less mapped reads in the end.

[ghd1998]

Thanks for your reply! Ambiguous reads are just assigned to the first best site so I don't think that is the reason, it appears that not all the reads are attempting to be mapped? When I change the ambiguous flag the number of reads being mapped doesn't change, only where the reads are assigned, any ideas?

[GenoMax]

How much memory are you assigning to this job? Have these reads been scanned/trimmed before splitting?

Have you also looked at the output of these reports?

scafstats=<file>    Write statistics on how many reads mapped to which scaffold to this file.
refstats=<file>     Write statistics on how many reads were assigned to which reference to this file.
                    Unmapped reads whose mate mapped to a reference are considered assigned and will be counted.

[Ram]

Is BBSplit sensitive to adapters/quality stats?