[darthsequencer]

Ahh... sorry, but nope That would require reading the file twice, so it would not be easy to implement.

Oh yeah - duh

[Brian Bushnell]

Ahh... sorry, but nope That would require reading the file twice, so it would not be easy to implement.

[darthsequencer]

Actually, while it's not documented, the flag "minlength" also works with BBMap. Reads shorter than that will be discarded completely (they won't be output as unmapped).

Thanks for the tip - save me some space so I don't have to make additional files. I deal with a lot of libraries with different read lengths.

Is there a way to make BBMap require the smallest read to be some fraction of the longest read length? I know that's a niche use but BBMap always suprises me with it's built in functions.

[Brian Bushnell]

Actually, while it's not documented, the flag "minlength" also works with BBMap. Reads shorter than that will be discarded completely (they won't be output as unmapped).

[GenoMax]

You could pre-filter input data with

reformat.sh in=file.fq out=filt.fq minlength=N

[darthsequencer]

Hi Brian,
Is there a way to set a minimum input fragment length when mapping with BBMap?

[Brian Bushnell]

To do that, you'd need to trim all soft-clipped bases. I don't have any programs that will do so, but it looks like I could add it to Reformat without too much difficulty.

[jweger1988]

Hi Brian,

Is there anyway to use bbmap (or any other of your tools) to map a read to a reference file and then trim anything to the left of the reference sequence?

For example

My reference is
XXXXXXXXXXXXXXXXXX

And my reads are

NNNNXXXXXXXXXXXXXXXXXX
NNNXXXXXXXXXXXXXXXXXX
NNXXXXXXXXXXXXXXXXXX

I want them to be
XXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXX

I basically want to just trim anything of the left of the reads that doesn't match my reference? Thanks in advance.

[sdriscoll]

Oh yes! Thanks Brian.

[Brian Bushnell]

You can use the flag "intronlen" to control this. For example, "intronlen=100" will change the reporting so that all deletions of at least 100bp will reported using 'N' instead of 'D' in the cigar string.

[sdriscoll]

I was testing something with BBMap today and I realized I had forgotten something about how to use it and couldn't figure it out. I was mapping RNA-Seq and I thought that it would report spliced alignment cigars (using sam 1.3) as /[0-9]+M[0-9]+N[0-9]+M/ but it was reporting the introns as deletions with the [D] cigar value. Is that right or is there a way to get it to not do that?

Thanks-

[Brian Bushnell]

Hi Jweger,

Sorry, I don't have anything that will do that. Clumpify allows something sort of similar; you can create consensus sequence from raw reads, and map those. But that loses per-base depth information which is important for variant-calling, so I don't think it's what you want.

[jweger1988]

Hi Brian,

I'm wondering if there is any way you can get bbmap (or another of your tools) to give a consensus sequence of an alignment? I went to map to a ref sequence and then use that to create a consensus sequence and then use that to map again and then call variants.

Thanks!

[Kristian]

Hi Brian,

Is this a Nextera long-mate pair library?

This is Nextera XT prep, and the adapters/linker were trimmed along with demultplexing directly through Illumina RTA pipeline. On my end the bbduk2 left and right trimming had almost no effect, and the splitnextera also had only a few hits (0.1%) to the adapter, so I'd guess the dataset is (for the most part) trimmed.

[Brian Bushnell]

Hi Kristian,

Is this a Nextera long-mate pair library? Those need special processing before they can be mapped. Or... can you give me any more information about the library construction, and the trimming methodology? The library has an extremely high error rate (particularly with read 2), less than half of the reads map to the mito, and it appears that both adapters and transposase are still present.... also, I'm measuring the median insert size as 159 (BBMap) or 133 (BBMerge), so there are a lot of pairs with insert size shorter than the sequenced read length; those might be displayed differently in IGV depending on whether the adapter portion was soft-clipped (which bwa would do by default) or not (bbmap does not soft-clip by default).

I adapter-trimmed the reads and error-corrected them, but still under 50% map. I'm not really sure what's wrong with the library. But, I don't see anything unusual about the pairing orientations. I get 45.5670% properly paired with "rcs=f" (require correct strand = false) and 45.5481% with "rcs=t", so only 0.02% map in the wrong orientation.