I personally haven't been very satisfied by the BreakDancer result from any of the data that I have tested. However, to be fair, I have also yet to find an alternative algorithm using a similar principle that works better (I like CoNIFER for indels, but I don't have a favorite tool for translocations or inversions).
Is it possible there was a problem with your simulation? It seems like it is had to distinguish a BreakDancer problem from a simulation problem at this point.
Assuming the simulation works, how many supporting reads are there for each break point? Maybe the coverage isn't high enough?
Also, are you providing BreakDancer with normal alignments as well as translocations? I know that it crashes with HaloPlex data because that is amplicon based and the paired-ends don't have any sort of consistent insert size (so, BreakDancer doesn't return anything). So, I know for certain BreakDancer won't work if you only provide the reads from the translocation.
Many thanks!!!
I have only translocation reads in my file. I did not include any normal reads. I used wgsim to create fastqs and did alignment using bwa with default parameters to get bam files.
However, the insert size distribution with my data shows a normal curve.
I will definitely create another file with the normal reads as well and try once again.
The 'translocation only' file worked perfectly with SVDetect though.
Thanks once again. I will update soon.
Many thanks!!! I have only translocation reads in my file. I did not include any normal reads. I used wgsim to create fastqs and did alignment using bwa with default parameters to get bam files. However, the insert size distribution with my data shows a normal curve. I will definitely create another file with the normal reads as well and try once again. The 'translocation only' file worked perfectly with SVDetect though. Thanks once again. I will update soon.