Use Phables on a mixed community

[bhagavadgitadu22]

I want to use Phables on stream biofilm metagenomes. I have a very intricate gfa graph obtained after MEGAHIT and during my tests Phables was able to solve only 4 case2_linear phages where several nodes are actually merged. I am wondering if I can tune some parameters to improve the performance?

I could also start from a different gfa graph. For instance in Hecatomb a Flye assembly of the contigs from my different samples is proposed: it is more doable than a cross-assembly in terms of resources and would simplify the gfa graph. I am afraid it would introduce chimeric contigs though.

[Vini2]

Hi @bhagavadgitadu22!

Thanks for your interest in Phables! Sorry about the delay in getting back to your issue.

In my experience, I haven't found MEGAHIT assemblies to work well with Phables. However, you can still try reducing parameters like --mincov to 1, --alpha to 1 and --covtol to 50. You can play around with these parameters to see if the results improve. You can find details of these parameters here.

I believe Hecatomb would be a better option to assemble as it will provide a more connected graph with longer contigs.