Hi everyone,
Briefly, I applied coverage-based binning method plus pentamer-frequency ordination on metagenomic assembled contigs, to isolate individual genomes.
Now I'm looking for clever (confident, efficient) methods to answer following questions:
1) Who do those genomes belong to? (ie. identify their taxonomy)
2) How complete are they?
For (1) I've tried AmphoraNet and found it quite good. But what about blasting everything against NCBI-genomes database? Would that work fast enough on a small server (Xeon, 16 cores, 96 GB memory)?
Question (2) could probably be answered if (1) is reached succesfully.
However, it could happen that I get, e.g., a novel Proteobacteria, but how to be sure this is really the case and it is not due to misassembly issues. It's a challenging discipline.
Thanks for your insights.
Briefly, I applied coverage-based binning method plus pentamer-frequency ordination on metagenomic assembled contigs, to isolate individual genomes.
Now I'm looking for clever (confident, efficient) methods to answer following questions:
1) Who do those genomes belong to? (ie. identify their taxonomy)
2) How complete are they?
For (1) I've tried AmphoraNet and found it quite good. But what about blasting everything against NCBI-genomes database? Would that work fast enough on a small server (Xeon, 16 cores, 96 GB memory)?
Question (2) could probably be answered if (1) is reached succesfully.
However, it could happen that I get, e.g., a novel Proteobacteria, but how to be sure this is really the case and it is not due to misassembly issues. It's a challenging discipline.
Thanks for your insights.
Comment