Determining the accuracy of a "merged" genome assembly

Question

[mmacd]

Hello all,

I have merged two large (~2.6 gb) genome assemblies from the same organism into one genome assembly using the GAA tool. I am looking for a way to determine how correct this new merged assembly is and to identify mis-assembled contigs/scaffolds. I have the contigs/scaffolds for all three assemblies now, as well as the raw reads from one of the original assemblies.

One idea could be to map the raw reads back to the new merged assembly and see how well this mapping is. However, I am not sure what mis-assembled regions would look like, i.e. low or high coverage?

Another idea is to align the contigs/scaffolds from the two original assemblies to the merged assemblies. I am currently using nucmer to do this and can visualize the alignments by mummerplot and in ACT. Two issues I am having, though, are what parameters to use in the alignment and how to automate this process.

Any help would be appreciated, thank you!

Answer 1

[Laura]

Have you tried aligning cdnas from your species to see how well they align?

Answer 2

[marina.v.yurieva]

You can combine all of the above :) Align two assemblies to the final one together with the reads and see where the differences are. Some reads might align better to one region in the first assembly while the others - to some different region in the second. So you'll have to take the best contigs from both.

Answer 3

[mmacd]

I will be trying what you suggested, marina.v.yurieva, thank you!

I have not tried cDNAs from my species yet, but I may try that as well as a last check.

[Vivek]

That is likely not a last check. Checking if the cDNAs are present in the draft assembly is one of the first QC steps.

Answer 4

[maheshhbg]

You can run QUAST (Quality Assessment Tool for Genome Assemblies) tool against merged assembly with individual assemblies.

It is very simple to operate and here is the link.

http://bioinf.spbau.ru/quast

Let me know, if you need further clarification on this.

Good luck !