Inquiry Regarding Delly's Performance in SV Genotyping Benchmark

[Cheng Quan]

Hi everyone, 

I am currently conducting a benchmark study on SV genotyping methods and have included Delly2 in my analysis. However, I found that Delly2 does not perform well on our simulated dataset, especially in terms of recall, precision and F1 score. I would like to know if there are any insights that could help me understand the potential causes of this situation, or suggestions on how to optimize its performance in this case.

To facilitate further discussion, I have made the simulated data, SV dataset, genotyping results, and pipeline scripts available on our website: https://www.omic.tech/ftp/public/test_script_for_genotyping/. Please feel free to review these data at your convenience. Any suggestions or guidance would be appreciated.

Many thanks!

Quan Cheng

Research Associate
Beijing Institute of Radiation Medicine  

[Tobias Rausch]

Hi Quan Cheng,

Delly is not a general purpose genotyper because the genotyping function in delly requires the local assembly of the SV allele (INFO/CONSENSUS and INFO/CT). In short, the input to '-v' should be a delly-generated VCF file. 

Best, Tobias

--
You received this message because you are subscribed to the Google Groups "delly-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to delly-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/delly-users/d25f42ef-69ce-47d6-94ff-5c82324f3200n%40googlegroups.com.