Extracting and scaffolding microbial genomes from complex communities

[MikeT]

Hi all,
For a recent project, we managed to sequence a microbial biofilm growing on a surface. It was a mixed community composed of both filamentous and single cells. We first sequenced the community as a whole, and then sequenced only the >10 um size fraction, which held almost only the filamentous portion.
I managed to extract 3 genomes from the >10 um size fraction by both applying MetaBAT and by manually selecting contigs that formed isolated clusters (after merging the whole community and size fractionated samples) with VizBin. The analysis of gene markers confirmed that they belonged to two different Proteobacterial classes and I managed to isolate the most closely related genome for each environmental genome.
Now I would like to scaffold my contigs and complete them as much as I can. Both samples were sequenced on an Illumina MiSeq platform with 2x300 bp paired-end reads. I tried by aligning the contigs on their most closely-related genome with CONTIGuator and then trying to close the gaps with GapCloser, but I'm not sure it's the best strategy.
What would you suggest me, instead?
Best regards

-MikeT

[fibar]

Hi Mike,
I'm interested in your post. As far as I know, what you say sounds good.
I haven't tried it yet, but maybe with anvi'o you can check where you are standing:

http://merenlab.org/projects/anvio/

[Markiyan]

Nextera/Lucigen Matepair data is strongly recommended for gap closing!

Dear Mike,

So well done for sequencing them with 2x300 run mode, but remember that maximum insert sizes that would cluster are still 1kb - 1.2kb, so in order to be able to jump over any repeats over 1kb (rRNA, etc) - the matepair library is a must (if you can get enough DNA (~2ug)).
I would do at least 1 nextera 3kb+ matepair library, making sure i give it enough DNA, run 2x300 or 2x250 mode,
than properly process the data and reassemble. If the range is still to short - you can go up to 15-20 kb if you use blue pipping or gel size selection.

After that, repeat the gap closing attempts.
You can also try pilon for that.

PS: if you have access - you can also give a Pacbio or 10K or Molleculo a try.
--
Best Regards,
Markiyan.

Hi all,
Both samples were sequenced on an Illumina MiSeq platform with 2x300 bp paired-end reads. I tried by aligning the contigs on their most closely-related genome with CONTIGuator and then trying to close the gaps with GapCloser, but I'm not sure it's the best strategy.
What would you suggest me, instead?
Best regards
-MikeT