Differences between .log and plasmid.fasta file about platon

[oschwengers]

Hi Thorsten,
I took a look at you results and I can confirm that even though contig00080 perfectly aligns to the reference plasmid, this is not a bug of the software as its RDS of -2.5 is below the required conservative threshold of 0.1. Of course this is not correct in terms of actually being a plasmid contig, but that is due to the underlying heuristic filter. So this is merely a wrong result of Platon's methodology but not a bug in terms of implementation.

The fact that CP016413 appears more often in the log file than in the sequences is related to this and expected.

You could execute Platon in sensitivity mode --mode sensitivity to include contig00080, however this will result in less specific predictions and will include a couple of false positives.

[tthye1]

Hi,
Thanks for clarification. I will test the different modes. But I wonder why not always include all contigs with an almost 100% identity to plasmids of the corresponding database to the results?

Best,

Thorsten

[oschwengers]

There are genomic islands that are frequently found on both chromosomes and plasmids, as for instance ICEs. Therefore, a plasmid hit alone is not sufficient for a contig classification as it could also be an ICE within a chromosome.