Poly-A/poly-T soft clipping

[roomfortwo]

Hello there, I'm using novoplasty to recover some arthropod mitochondrial genomes from RNA-seq (quite trending these days) and I'm encountering that novoplasty tends to extend the poly-A tails from the reads (or poly-T in the 3'). Example below.

>Contig01+195792442 TTTTTTTTTTTTTTTTTTTTTTTTTTATCTAGATACAACTTCCAATATATCTACTTTGTTACGACTTATCTCTTTTTTCAGAGAGAGCGACGGGCGATATGTACATAATCTAGCCCTAATTCATTAGAATAAATTAATATCTAATTACAT ... CAACATAATTTTCCTCCAGCCGATCATACGTACAATCAATTAAGTGTTTCATCTTTATAAAAAAAAAAAAAAAAAAAAA

I tried to remove the polyA/polyT sections from the reads and failed mainly because trimmers tend to remove AT-rich biological sequences of interest. I was wondering if there's any flag I could implement in which I could either soft clip the polyA extensions or just ignore them in order to retrieve genomic seqs.

Thanks in advance, Luis AF.

[ndierckx]

Hi, I don't have experience with RNA-seq data, is the complete mitogenome covered by it, or just certain areas? Does it only extend with poly A/T at the ends when no further extension is possible , or you think further extension is possible? You can run with extended log to 1 and send me that file, can have a look

[roomfortwo]

Hi,

1. I think the coverage of the mitogenome depends mainly on the RNA extraction methods (if you select mRNA based on poly-A tails for example) and random factors (like the expression of the mtDNA itself, mtDNA contamination),
2. I think further extension is possible, but the reads to extend might be less abundant than the reads with polyadenilation, that's why I first thought of clipping the polyT/polyA.

Here you have three logs from assemblies of the same library (SRR3458647), different seeds, the RNA from this library was selected for poly-A tails.

log_extended_hans-3-k29.txt log_extended_node49-k29.txt log_extended_scu-b-k29.txt

This in the other hand is a log from a whole RNA sequencing library (I have a bigger logs but they are too heavy), with this library I assembled the whole mtDNA. log_extended_mtDNA-frl2015-39.txt

Let me know what you think, cheers,

[ndierckx]

How long are those mitochondrial genomes you try to assemble?

And the example of your first comment is not one of those 3 you send I guess?

I don't think it is always polyA tail problem, the coverage seems to go to almost 0 before the polyA starts, so not sure if it can extend further And I think some are just A-repeats, looks like the data has a strong error rate and reduced coverage around these regions.

If you have a specific sequence or polA you want me to check i will do, but I can't check all files completly

[roomfortwo]

Hey, I'm trying to assemble 14k genomes and the "strong error rate" you mention I believe are the NUMTs. If you think Novoplasty it's not extending due to low coverage and the strong variation, I'll take it.

Thank you so much for you insight!

[ndierckx]

Hi, NUMTs are not the problem, I usually doesn't affect the assembly. I mean strong error rate around single nucleotide repeats, depends on the library preparation and illumina kit you used.