Limits of Indel Detection

[mitchum20]

Dear all,

I am wondering if there is a maximum possible indel length, which is still detectable in an alignment of paired-end 150bp Illumina reads.

I am using the BCFtools package to call Indels on novoaligned short reads and noticed that called indels dont exceed a certain number like 70bp.

Thank you for your comments.

[Meyana]

Might be a limit to the indel caller.

I'd suggest using bbmap for mapping and CallVariants (also in the BBtools suite), I'm successfully calling long (>10kb) indels using 150bp paired end reads using these tools.

[mitchum20]

Thanks your suggestion. Can you suggest me a handbook or guide for this method ?

[Meyana]

Please refer to the bbtools manual here:

https://jgi.doe.gov/data-and-tools/bbtools/bb-tools-user-guide/

And the callvariants page here:

https://github.com/abiswas-odu/Disco/blob/master/bbmap/callvariants.sh#L119

You could likely just follow the recommended/default parameters and see where that gets you first. Then, if there's something you're not picking up (I like to compare with how the data looks in IGV) try fiddling with the parameters.

Please also take a look at the BBMAP thread on this forum (lots of goodies!) or this thread by GenoMax (http://seqanswers.com/forums/showthread.php?t=58221), but feel free to ask if you have something more specific

Only thing I would like to point out: For my data, the kmer value set during mapping influences indel calling a lot. Please use the default setting first, in my experience lowering it did bad things to my indel calling....