Binning finishes successfully but with errors about metabinner HOT 6 CLOSED

Hi, Chris,

Could you please tell me the number of samples of your dataset? It seems that there are no quality bins generated using the coverage information only, which makes it fail to finish the second-stage ensemble process. As stated in the manuscript (https://www.biorxiv.org/content/10.1101/2021.07.25.453671v1.full.pdf), some component binning results for integration are generated using coverage information alone as features, and we recommend applying MetaBinner to multi-sample datasets.
MetaWRAP is a good choice for small-scale datasets or datasets with few samples. Or maybe the ensemble results generated by the first-stage ensemble process of MetaBinner using the combined features is fine. You can find them in "../metabinner_res/ensemble_res/X_t_logtrans_2postprocess/greedy_cont_weight_3_mincomp_50.0_maxcont_15.0_bins".

Thanks for your report, and we will improve the error reminder. Please feel free to let us know if you have any further questions.

Best wishes,
Ziye

from metabinner.

Hi Ziye,

Thanks for the quick reply! It was only one little mock community I was messing around with, so that will explain it then.

Thanks again for the help!

Chris

from metabinner.

Hey, I'm running into the same problem, but was wondering: is there a minimum sample size you recommend? Or a lower limit beyond which you recommend switching to a program/pipeline?

from metabinner.

Hi. On my side MetaBinner performs very poorly on low diversity metagenomes.
Additionally to previous comments, how do you create multi-samples dataset? Are we supposed to merge bamfiles ? This is very painful to do.

from metabinner.

Hey, I'm running into the same problem, but was wondering: is there a minimum sample size you recommend? Or a lower limit beyond which you recommend switching to a program/pipeline?

Hi,

I recommend running MetaBinner on the datasets with no less than ten samples, but there isn't an apparent minimum sample size. The performance may be influenced by the sample size, the datasets' complexity, the assembly's quality (e.g. assembly contiguity), and so on. Sorry for not replying in time.

Best,
Ziye

from metabinner.

Hi. On my side MetaBinner performs very poorly on low diversity metagenomes.
Additionally to previous comments, how do you create multi-samples dataset? Are we supposed to merge bamfiles ? This is very painful to do.

Hi,

The binner was not developed to handle the low diversity metagenomes; understandably, it didn't perform well on the low diversity metagenomes. But we hope that MetaBinner will also perform well on such datasets. If the low diversity metagenomes you used are publicly available, could you please send us the link to the metagenomes and let us figure it out. Thanks very much.

We align the reads from each sequencing sample against the assembly file to generate multiple bam files, calculate depth (coverage) for each bam file and merge the outputs. The process is similar to that of most binners.

Best,
Ziye

from metabinner.