google / deepconsensus

DeepConsensus uses gap-aware sequence transformers to correct errors in Pacific Biosciences (PacBio) Circular Consensus Sequencing (CCS) data.
BSD 3-Clause "New" or "Revised" License
229 stars 36 forks source link

deepconsensus on metagenomic data #44

Closed ay-amityadav closed 2 years ago

ay-amityadav commented 2 years ago

Hi,

Is it ok to run deepconsensus on HiFi metagenomic data? I can see that it has been tested on human genomic data.

Many thanks

AndrewCarroll commented 2 years ago

Hi @ay-amityadav,

Thank you for your question. DeepConsensus should work for metagenomic data. We've tested DeepConsensus on E.coli and several plant species. For E.coli we see higher accuracy as measured against the known sequence of the reference strain. For plant species, we see improvements to assembly contiguity and correctness.

As a result of those investigations, we believe DeepConsensus should work well for metagenomic samples as well.

ay-amityadav commented 2 years ago

Thank you very much!