search

clemgoub / dnaPipeTE

dnaPipeTE (for de-novo assembly & annotation Pipeline for Transposable Elements), is a pipeline designed to find, annotate and quantify Transposable Elements in small samples of NGS datasets. It is very useful to quantify the proportion of TEs in newly sequenced genomes since it does not require genome assembly and works on small datasets (< 1X).
48 stars 11 forks source link

Discrepancy between Trinity.fasta and Reads_per_component_and_annotation output #87

Open emiya52 opened 1 week ago

emiya52 commented 1 week ago

Hi,

I am recent graduate student who is doing raw dna sequence analyzing from the Red Sunflower Seed Weevil. In this research, I am trying to take an approach of utilizing both dnaPipeTE & deepTE(CNN based classification tool) hoping for better results in TE classification.

The approach I take is as follows:

When taking this approach, I figured that there is a discrepancy between the number of entries(rows) of the Trinity.fasta data set and the reads_per_component_and_annotation output, which the latter lacks about 4k entries.

I was hoping to have a better understanding on why this discrepancy exists. My guess is that within the dnaPipeTE pipeline, after Trinity is done configuring the sequence the data is handed over to Repeat Masker for annotation and quantification, and there is some filtering done by RM in this process that takes out certain reads that don't meet a certain threshold, but not sure whether it is true or not, since the final output file(reads per component and annotation) also includes unknown elements.

I am hoping if you could confirm or provide any insights regarding the discrepancy between the Trinity.fasta output file and the reads_per_component_and_annotation file, or if there could be an alternative output file that I could be utilizing for the approach I am taking.

Thank you for reading this, and thank you for developing and sharing such a wonderful tool.

with best regards,

clemgoub commented 1 week ago

Hello! Thanks for reaching out.

It is most likely that the 4k contigs that have disappeared between the file Trinity.fasta and reads_per_component_and_annotation came from low-copy repeats (possibly even non-TE sequences). This happens, because the read sample used for quantification is drawn independently that the sample(s) used to assemble the repeats with Trinity.

By default Trinity does 2 iterations with 2 independent samples, while a third independent sample is used for the quantification.

So unless there is something else you noticed, I think this is all normal based on the sampling strategy of dnaPipeTE.

Cheers,

Clément