Hello everyone,
We are doing epi-transcriptomic (m6A), polyA length tail analyses using nanopore DRS dataset. We have 3 replicates of each condition. The sequencing facility used same flowcell to run 3 replicates of same condition sequentially after washing them with nucleases after running each replicate (is this common practice??). So, as we lose some pores after each run, we have a gradual decrease in read quality and length in the 2nd and 3rd rep for each condition. After we basecalled the reads, we can see the median read length is almost ½ times in 2nd and 3rd rep compared to 1st rep.
Now we are trying to figure out upto what extent this difference in read distribution between replicates will affect our analyses.
We tried using stringent quality (Q>=9) and length cutoffs for our reads (>200 nts), it does improve the overall read distribution but then we will miss the reads which corresponds to shorter original transcripts.
As most of the m6A sites and polyA tail tend to present in the 3’ end of the reads and the nanopore sequences the reads from 3’ end only, we might not get much affected.
If anyone have any suggestions we would appreciate it.
Just an anecdote as a heavy nanopore user - we wash and reuse flowcells for DNA samples all the time but never reuse after RNA. ONT has an official protocol for washing. The newer flow cells (which I don't have much experience with) may handle washes better than the old (which I have a lot of experience with).
I assume you mean there will be a 3' bias
but the sequencing won't start from 3' end.Huh? ONT DRS sequences from 3' to 5'
Thanks. Was not aware of this since I have not worked with such data. Checked on ONT direct RNA kit and it looks like the basecalling software automatically flips and displays the data as 5'-3' though sequencing starts from 3'-end. So 3' bias will be present in the data.