Dorado for m6A modification detection

Question

[baibhu1234]

Hello everyone,

I am working with plant Nanopore DRS data based on RNA004 chemistry, and trying to use Dorado for m6A calling in my DRS data.

I do not have any access to GPU but i have a HPC, and i have ~15 million reads/sample, can that be basecalled with base modification in 48 hours time limit?

(I am getting the segmentation error repeatedly, i will have to figure out a way to handle that)

also,

(1) once i get the bam output file from dorado, how do i get the genomic/transcriptomic location

(2) How to proceed to the differential modification analysis between conditions?

Answer 1

[GenoMax]

Prior thread for reference: m6A modification prediction using Nanopore DRS data (RNA004)

I am getting the segmentation error repeatedly

There is no way around this. Ask the sequence provider to run the analysis for you if you are not able to do this locally.

I have ~15 million reads/sample, can that be basecalled with base modification in 48 hours time limit?

Difficult to say. We don't know what kind of HPC you have and what is the median read length you have. You could run multiple jobs with sets of POD5 files to parallelize the process. It will also depend on the model you wish to use. With "super accuracy" ~15 M reads could take 3-4 days even on a single GPU. "high accuracy" calls should take about half that time.

once i get the bam output file from dorado, how do i get the genomic/transcriptomic location

You will need to align the data to the reference genome. You can easily convert the unaligned BAM to fastq.

How to proceed to the differential modification analysis between conditions?

ONT makes a program available to do this called modkit (LINK). Use that.