Genome coverage of MAGs while submitting to NCBI

Question

[Konstantin]

Hello everyone. I am trying to submit my metagenome assembled genomes to NCBI through its "genome submission portal". It asks me about the genome coverage of my genome. What exactly should I specify, considering I submitting MAGs? I have no expected genome size for them. Should I map raw reads from th bulk metagenome to my MAG's assemblies? If yes, what exactly will I get from the results of this mapping, except of knowing what portion of reads in the dataset is my MAGs?

Thank you very much in advance, Konstantin.

Answer 1

[Mensur Dlakic]

Genome sequencing coverage (AKA, sequencing depth) is the average number of reads that map to a given genome. It tells you how many different times a genome has been sequence (on average). Generally speaking, deeper coverage means more reliable result, although that relationship may break down at really deep coverage.

For metagenomes, genome coverage is related to the population abundance of a given MAG. More abundant organisms will have their DNA present in larger amounts after extraction, and will be sequenced more on average.

You might want to try to some Googling on this subject:

https://www.google.com/search?q=genome+coverage+meaning

[Konstantin]

I now understand what this metric is in the context of metagenomics, thank you very much.

But immediately another question appears. How much sense does this metric have for genomes assembled from very diverse metagenomes, sush as that of a soil? It is hard to find a microbe in such metagenome that would have been sequenced remarkably many times. So can we even speak about reliability here, if all such MAGs are, on average, not very reliable. We just operate very small numbers here. Maybe 1 or 2 percent for the most common genome in a sample that was sequenced. Is it significantly more reliable than a MAG with the coverage estimation of 0.1 - 0.5 %? How do you think?

And again, I appreciate your help!

[Mensur Dlakic]

A short answer to your question is yes, at least when done properly. The rest I leave to you to research.

It is customary to upvote and/or accept the answer if it solved your problem.

[wn835166087]

Hi, I have a follow up question about this. During my binning process, I obtained the coverage of each contig. Can I take the average coverage of the contigs that belongs to one bin as the coverage of the MAG? Any insights would be very appreciated. Thank you! Best regards, Nan

[Mensur Dlakic]

That should be fine.