Hello everyone,
Concoct 's result in Metawrap binning module has only file (unbinned.fa). Maxbin2 and metabat2 outfile has a lot of bins, so is concoct's binning result normal? Thank you in advance.
It is not normal, but it indicates that CONCOCT couldn't create any bins. That's the whole problem with most of these binning programs - they tend to be black boxes. In other words, they do not output the clustering solution in a way that is easy to understand and inspect visually. I'd make sure that this is not caused by the minContig size cutoff, which for metabat2 is 2.5 kB. I don't know if CONCOCT even applies this cutoff. If not, it could be that small contigs are creating so much noise that there is no clear bin separation.
There are algorithmic differences that will cause one binner or the other to perform better on a given dataset, and small differences are to be expected. However, this is not a small difference. That's why I'd be weary of the other two solutions as well. If a competent and well-tested program such as CONCOCT can't find a clear separation to bin sequences, that indicates a tough dataset and should be happening with other programs as well. It may be worth setting the minContig at 2 kB for all of them, or if not possible (not sure if CONCOCT has that option) then manually remove all the contigs smaller than 2 kB before feeding into CONCOCT.
seqtk seq -L 2000 my_contigs.fa > contigs_2000_and larger.fa
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Dlakic, thank you. CONCOCT will cut longer contigs into 10 - 20 kb pieces prior to mapping, and contigs shorter than 1000 will not be included for analysis. So, whether so much noise was produced by small contigs may not main reason?
I already made a suggestion, and there is only one way to find out if it will help CONCOT create some bins. For some datasets a minContig cutoff at 1 kB is enough, and for the others it needs to be higher than that. To truly compare what metabat2 and CONCOCT are doing and the bins obtained, the minContig cutoff needs to be the same. It is 2.5 kB in metabat2 by default, which I consider a better choice than 1 kB.
All right, Thanks for your suggestion. I will try to set the same minContig cutoff between metabat2 and CONCOCT. Wish have a good result.