Running purge_dups on a hybrid assembly (optical mapping + reads)

Question

[pablo]

Hello,

I have a hybrid assembly I obtained with Pacbio assembly and an optical map from Bionano. This assembly still remains rather heterozygous, that's what I would like to run pourge_dups to remove haplotigs on this assembly.

I can run this tool on the primary assembly I obtained with my Pacbio hifi reads , but is this tool suitable with a hybrid assembly (Pacbio hifi reads + BioNano optical map) ?

Best

[lieven.sterck]

BioNano optical maps do not add/alter sequence in your assembly if I remember correctly, so I don't see any harm/issues with running it on the hybrid assembly.

[pablo]

Thanks for the tip, I'll try that.

[pablo]

I ran purge_dups on my hybrid assembly. I guess it is still rather heterozygous (1.9Gb while I expect 1.3Gb genome size) but purge_dups can't find any haplotigs. The dups.bed file is empty and so, the purged assembly is the same than my initial assembly. Is that possible you think?

[lieven.sterck]

yes, that is possible. Likely depends on the criteria used by purge_dups, check the manual/help to see what it uses.

You might need to tweak the parameter settings a little.

[colindaven]

I would say that is/could be more of problem with your fasta headers. I tried purge_dups on 4 assemblies and it failed to detect anything on 3, or only regarded the first contig. These 3 assemblies all contained hyphens "-" in the contig fasta headers.

A fourth assembly ran fine, as it did not contain complex fasta headers (eg sample_contig1 worked fine).

I then pruned the complex fasta headers of the first 3 and resultant bed files were generated for all.