search

glogsdon1 / sunk-based_assembly

14 stars 2 forks source link

genome-wide sunk-based assembly #4

Open dcopetti opened 3 years ago

dcopetti commented 3 years ago

Hello,

Thank you for sharing the code for this very clever method to resolve complex regions. Would you envisage to use this method genome-wide? I mean, not having the need of a parent read next to a gap (just a set or random reads maybe?) Thanks,

Dario

glogsdon1 commented 3 years ago

We're hoping to develop something like that, and I know other teams are working on this too. If we can get an assembler that does this genome-wide, it would be a game-changer.

dcopetti commented 3 years ago

I agree! I have an idea based on a similar concept, I see the benefits that such a method would bring, especially in heterozygous or polyploid genomes. Where is the bottleneck in your method?

glogsdon1 commented 3 years ago

Right now, our bottleneck is mapping the SUNKs onto the ONT reads and stitching them together accurately, but we're working on making that part of the pipeline more efficient! I'm interested to hear your idea as well.

dcopetti commented 3 years ago

Our idea starts from the same concept but we would like to translate the edges between reads (given by the shared SUNKs) into the assembly graph directly. Then, existing tools should be able to turn the graph into a consensus sequence (to be polished). But I agree with you, mapping the k-mers efficiently is not an easy task and our method is not efficient at that yet. let's see who gets there first! (you!) ;-) Dario