Determine number and length of plasmids in bacterial genome (using something like Prokka)

Question: Determine number and length of plasmids in bacterial genome (using something like Prokka)

15 months ago by

suzuBell • 60 wrote:

I have been finding the Prokka software to be very useful annotating an assembled bacterial genome (contigs). I wanted to know if anyone has recommendations on how to determine if the assembled bacterial genome (contigs) contains a plasmid. If so, how many plasmids and length of plasmids.

I tried using PLSDB and PlasmidFinder with a bacteria genome I downloaded that reportedly has one plasmid - and both software returned no hits. I also ran these software on my assembled bacterial genome and received no hits (although I am not too confident in this result given that they returned no hits for the known downloaded bacterial genome, and it could be because I am not using them correctly).

My output from Prokka does not seem to contain information about presence/absence of plasmids and it seems it may not be designed to find plasmids. Any advice on an elegant solution to find plasmids in the assembled contigs (I do not want to do something too manual like blast each contig individually - because I plan to repeat the procedure for many assembled genomes)?

prokka bacteria plasmid • 661 views

ADD COMMENT • link •

modified 15 months ago by hafiz.talhamalik • 250 • written 15 months ago by suzuBell • 60

15 months ago by

Mark • 860

Mark • 860 wrote:

How did you assemble your genome? This is key to actually detecting plasmids. You might want to use specialised software for the assemble of plasmids such as plasmidSPAdes: https://www.ncbi.nlm.nih.gov/pubmed/27466620 Then use the PLSDB you mentioned to search the database.

Once you find a hit, you can use that reference genome(s) to annotation the plasmid(s) using Prokka. Prokka can annotate plasmids, you just have to point it in the right direction by providing it a good reference file. Follow his instructions on how to annotate plasmids.

ADD COMMENT • link written 15 months ago by Mark • 860

15 months ago by

Joe ♦ 18k

United Kingdom

Joe ♦ 18k wrote:

Prokka won't tell you anything about plasmids, and as you say, isn't intended to. The only way in which it might is if you can find a contig prokka has annotated that looks like it has a lot of 'typical plasmid stuff on it', e.g. phage elements etc.

However, prokka is deliberately conservative in its annotations. This is especially a problem for plasmid annotations which frequently have a lot of uncharacterised proteins with little to no experimental evidence.

Plasmid detection is still largely an unsolved problem in bioinformatics, its not all that surprising that the tools you've tried so far often fail to detect them.

Your best bet will be to find a good reference as the other answer alludes to, but I would advise trying several more plasmid detection tools and trying to get some sort of a consensus. A couple that spring to mind are PlasFlow, pATLAS, HyAsP, PlasmidSeeker.

Unfortunately this is generally one of those things where you just have to 'do the best you can'.

ADD COMMENT • link written 15 months ago by Joe ♦ 18k

15 months ago by

hafiz.talhamalik • 250

Pakistan

hafiz.talhamalik • 250 wrote:

There might be a possibility that your assembly is not good. So try checking that too.. or you can try reference based assembly too for good assembly results. As suggested earlier try plasmidSPAdes for assembling plasmid. another recent tool is HyAsP

ADD COMMENT • link written 15 months ago by hafiz.talhamalik • 250

Please log in to add an answer.

Similar posts • Search »