Brief note that we just released rastair v2.0.0 - ...
# methylseq
b
Brief note that we just released rastair v2.0.0 - this is a major version upgrade and not 100% command-line identical to rastair 0.8 which is used in methylseq. However, the differences are very minor so it should be fairly easy to adapt. The key difference between v0.8 and 2.0 is much more powerful variant calling. Rastair2 now features a full-fledged variant caller under the hood, which will not only identify SNPs that affect CpGs and adjust the estimated beta accordingly, but also identify SNPs that create a CpG and call methylation there. This adds several hundred thousand additional methylated positions in an average human genome! You can also get full-fledged vcf output in addition to the per-position and per-read bed formats as before. We have tested Rastair2 against both TAPS+ as well as Illumina's 5-Base data, and I'm happy to say that it works extremely well on both: our SNP calling accuracy is as good or better than Illumina's DRAGEN pipeline on 5-base data, and nearly as good as GATK's haplotypecaller on WGS data, while being orders of magnitude faster. For some preliminary results, see https://www.rastair.com - a more comprehensive pre-print is in the works. It would be great to get some feedback from the community and hopefully upgrade nf-core to include that new version. I'm not (yet) very familiar with how nf-core pipelines are put together, but I'd happily consult on what needs changing to migrate from 0.8 to 1.0. If you just want to test stuff out for now, the docker build of v2.0.0 is available, and we're also nearly there with a bioconda recipe. There's also binaries for Linux & Mac, see the website above for details.
🎉 3
n
Hej Benjamin! I was wondering, are there any plans in the future to include CHH and CHG calling in rastair (for plant methylation)? Or does the tool already handle these as well?
b
Hi Naomi! I got a similar question recently from someone else via the issue tracker. We would like to support CHH and CHG natively, but we're lacking test data. I was hoping someone could share some sequencing runs with us to use for development, but it hasn't worked out just yet
We do extract C>T variants at CHH and CHG, and report their strand bias, but only if you run
call
with
--all
in
vcf
mode. It's not the most convenient way of getting the info, I know
@Naomi Keehnen let me know if this makes sense?
n
I will test it out soon! 🙂 I will let you know how it goes
👍 1