How to assess the quality of metagenomics assembly?

Question

[beloved2012]

Hello everyone! I'm a beginner of metagenomics research. Presently we use soapdenovo package to perform metagenomics assembly then choose the optimal Scaffold according to N50 length (the longest one). I'm wondering that how should we assess the quality of metagenomics assembly except according to the N50 length? Thanks very much!

[beloved2012]

OK, thanks all of you. I'll try QUAST later.

Answer 1

[Brian Bushnell]

Quast is, indeed, very useful. However, it is most useful when you have a reference, where it can calculate the number of misassemblies and error rates. For metagenomes, another crucial metric - arguably, the most important metric - is the percentage of reads that map back to the assembly. If only 50% of your reads map to the assembly... it is not very complete. But if 95% of your reads map to the assembly, then even if it is somewhat fragmented, that's probably very good.

[rens.holmer]

It might also help to look at the percentage properly paired reads to detect any chimeras, something that seems especially relevant in a metagenome assembly.

Answer 2

[marina.v.yurieva]

Try quast

It gives you a report with the quality of your assembly.