Run Delly by chromosome

Question

[tg25]

Dear All,

How do you run Delly by chromosome? I have looked at the Delly GitHub page and I cannot find the answer. I have also looked at old questions like this Delly question but this question does not address how to run by chromosome.

I am running

delly call -t DEL -g ref.fa -o dellyOut.bcf file1.markedDup.bam file2.markedDup.bam file3.markedDup.bam file4.markedDup.bam . . . file100.markedDup.bam

For each of the structural variations I am running the options below:

-t DEL = Deletions

-t INS = Insertions

-t INV = Inversions

-t DUP = Tandem Duplications

-t TRA = Inter-chromosomal translocations

While this is okay, I would also like to run Delly by chromosome. If any of you know the list of options for Delly or where they are stated can you please post a link. I found the information about option -t by reading this Biostars question.

Thank you for your help.

[Vitis]

Can you split the BAM into chromosomes and run Delly? You may split BAM using samtools or sambamba.

[tg25]

Yes, I can split the BAMs by chromosome. With doing this I have some questions:

1. If I split the BAMs then realign with bwa, would Delly detect inter chromosomal translocations?
2. Or if I split my already bwa aligned BAMs, would Delly detect inter chromosomal translocations?

Before using Delly, I used BreakDancer which has an option -o : STRING operate on a single chromosome [all chromosome]. If Delly as a similar option this would be great.

Additional Information: I forgot to add that BreakDancer has an option -t : only detect transchromosomal rearrangement that I use to detect inter chromosomal translocations. Because of the number of files that I work with I found that splitting the analysis by chromosome and then running option -t for ctx on the same (unspilt bams) worked really well in breakdancer. All I have to do is then merge the outputs. My analysis was timing out in Breakdancer with splitting the analysis fixed this and I would like to do the same with Delly.

[Vitis]

Translocation callers usually use split read pairs or reads across chromosomes (secondary alignments, SA tag), so it does not make sense to call translocations using single chromosomes, unless you know the chromosomes on which translocation happened. In that case, you may slice the BAM for the two chromosomes only.