I am performing whole-genome bisulfite sequencing (WGBS) analysis on a plant species for which no reference genome is available. To proceed, I used the genome of a closely related species as a reference; however, this resulted in very low mapping efficiency when using Bismark. I also attempted to relax the alignment stringency using parameters such as --score_min L,0,-0.6, and its still running.
However, I would like to assess whether my sequencing reads contain potential contamination, such as bacterial sequences. For this purpose, I am considering a de novo assembly approach. However, since the data consists of bisulfite-treated reads (where unmethylated cytosines are converted), I am uncertain about the suitability of standard de novo assemblers.
I have come across a tool called epiGBS, but I understand that it is primarily designed for RRBS (Reduced Representation Bisulfite Sequencing) data.Would it be appropriate to use tools like SPAdes for assembling bisulfite sequencing reads, or are there more suitable approaches or tools specifically designed for handling bisulfite-treated data?.
1 answer
This is really strange. I have never heard of de novo assembly attempts of bisulfite. I'd expect it would fail completely given the biased sequence. In any rate, Illumina assemblies are rather 2010 and probably not publishable in the T2T era.
It sounds like you need a reference genome to do your study. Quite how the person who designed the experiment expected it to actually work without a reference is completely beyond me.
You could - and probably just should - get a decent reference genome using a good service provider using an ONT promethion flowcell (depending on genome size). Then use hifiasm to assemble or shasta. You'll need at least 20-40x coverage per haplotype.
Good luck.
Log in to answer this question.