search

Magdoll / MagPhase

Phasing for metagenomics using PacBio long reads
BSD 3-Clause Clear License
17 stars 3 forks source link

Is it compatible with nanopore reads? #1

Open cdiazmun opened 1 year ago

cdiazmun commented 1 year ago

Dear,

First of all, sorry for asking about nanopore in the first issue when it is clearly indicated that this pipeline is ment for PacBio HiFi reads.

BUT, since it takes as input the assembly and the BAM file, I guess it could run also with an assembly and alignment file created using nanopore reads. So, my question is, do you expect similar efficiency at separating haplotypes within MAGs? Because, a priori, the biggest difference is the lower accuracy of nanopore reads.

As you can imagine, I have some MAGs in my dataset generated using nanopore's long read sequencing and it would be nice to test your approach on these.

Regards, Cristian

njdbickhart commented 1 year ago

Hello Cristian,

I can't answer on behalf of the developer, but I believe that read quality is the most important attribute here. Basically, the MAGPhase algorithm works when long reads have sufficiently low error rates (<= 1%) to enable accurate SNP calling.

I highly encourage you to test this with your dataset! The great thing about MAGPhase is that it was developed to enable the visualization of reads using IGV to confirm haplotype attribution. This makes it so much easier to validate results and you don't have to rely on "black-box metrics" to see if your dataset generated useable results.

Thanks, Derek

cdiazmun commented 1 year ago

Dear Derek,

Thanks for your answer!

I will definetly give it a try then. If I observe something worth mentioning, I will come back to you.

Regards, Cristian