Lightnews — Scholar-powered news

Dmitry Antipov

@dantipov.bsky.social

120 followers 48 following 23 posts

Genome assemblers and their friends. Postdoc at https://genomeinformatics.github.io/

genomeinformatics.github.io

Posts Media Videos Starter Packs

Dmitry Antipov @dantipov.bsky.social · Jun 27

Molecular Mechanisms in Evolution GRC conference short recap

Dmitry Antipov @dantipov.bsky.social · Jun 12

Verkko2 paper is finally published!
genome.cshlp.org/content/earl...

Verkko2 integrates proximity-ligation data with long-read De Bruijn graphs for efficient telomere-to-telomere genome assembly, phasing, and scaffolding

An international, peer-reviewed genome sciences journal featuring outstanding original research that offers novel insights into the biology of all organisms

genome.cshlp.org

Dmitry Antipov @dantipov.bsky.social · Feb 5

I believe this to be one of the most important things now. www.theguardian.com/us-news/2024...

Guardian missed that this guy is also a long-time business partner of Peter Thiel and is really influential among those tech bros.

I've read his manifestos. They are frightening and really match the news

He’s anti-democracy and pro-Trump: the obscure ‘dark enlightenment’ blogger influencing the next US administration

Key figures in the incoming administration follow Curtis Yarvin, who’s pushing for an autocratic takeover of the US

www.theguardian.com

Dmitry Antipov @dantipov.bsky.social · Dec 27

Verkko2 paper is finally on the biorxiv
www.biorxiv.org/content/10.1...

Verkko2: Integrating proximity ligation data with long-read De Bruijn graphs for efficient telomere-to-telomere genome assembly, phasing, and scaffolding

The Telomere-to-Telomere Consortium recently finished the first truly complete sequence of a human genome. To resolve the most complex repeats, this project relied on the semi-manual combination of lo...

www.biorxiv.org

Dmitry Antipov @dantipov.bsky.social · Nov 19

In the average lab you develop a proof of concept level software to support your part of a research project and do it alone. So realistically only the version control would be clearly beneficial.

The thing is that even computational methods development labs often miss a lot of your points

Dmitry Antipov @dantipov.bsky.social · Nov 19

This post hurts a lot!

Dmitry Antipov @dantipov.bsky.social · Aug 31

Finally on some samples we have all 46 human chromosomes T2T. Update to verkko 2.2, last big release of this summer! github.com/marbl/verkko

GitHub - marbl/verkko: Telomere-to-telomere assembly of accurate long reads (PacBio HiFi, Oxford Nanopore Duplex, HERRO corrected Oxford Nanopore Simplex) and Oxford Nanopore ultra-long reads.

Telomere-to-telomere assembly of accurate long reads (PacBio HiFi, Oxford Nanopore Duplex, HERRO corrected Oxford Nanopore Simplex) and Oxford Nanopore ultra-long reads. - marbl/verkko

Dmitry Antipov @dantipov.bsky.social · Mar 14

TLDR:
Prolonged reviewing process may lead to situations when at least a part of a paper is clearly outdated and looks weird at the moment of publication.
And I do not know how this should be fixed, assuming that we cannot always make review iterations fast.

Dmitry Antipov @dantipov.bsky.social · Mar 14

As a reader I would be happy if the journal would restart the reviewing process from scratch, with new reviewers. For all papers in progress, submitted too long (a year?) ago.
But as an author, I would be completely pissed off in such a situation.

Dmitry Antipov @dantipov.bsky.social · Mar 14

Likely reviewers had some other questions and on the next round had no reasons to return to sections they already found to be OK. And authors were not encouraged to provide unrequested changes.

Dmitry Antipov @dantipov.bsky.social · Mar 14

This is the answer. At September 2022 Verkko was already released and even preprint was published peer-reviewed; hifiasm was in process of adding the support for ONT.. No peer-reviewed publications for ONT+HiFi for both.

Dmitry Antipov @dantipov.bsky.social · Mar 14

So, one of two likely most popular tools in the field is completely ignored, and the other is run in the way it would produce clearly inferior assemblies! With no motivation in the text!

Junk paper? Bad reviewing quality?
None of the above!

Dmitry Antipov @dantipov.bsky.social · Mar 14

Ok, I can be biased about verkko, let's check hifiasm... Wow, it was run without ONT!

Dmitry Antipov @dantipov.bsky.social · Mar 14

Merfin is not an assembler, but I can understand what they mean. However, don't you feel that something is missing?

Dmitry Antipov @dantipov.bsky.social · Mar 14

So, it is an Illumina+ONT based polishing pipeline, that is focused on diploid genomes and assumed to work after Flye assembler. Stated to be comparable or even better than state of the art HiFi "haplotype-resolved" assemblers. Pipeline itself looks decent, so I go directly to benchmarking..

Dmitry Antipov @dantipov.bsky.social · Mar 13

Want to discuss a thing about reviewing process. It is far from "ai-generated alpha male rat" level of weirdness, but still weird for me, and I really do not know how it should be fixed.
Let's look on the article on diploid aware polisher. Published just a week ago
www.nature.com/articles/s41...

Dmitry Antipov @dantipov.bsky.social · Feb 20

Was not good in finding appropriate words. Of course it is not just "death", it is murder.

Dmitry Antipov @dantipov.bsky.social · Feb 16

Alexey Navalny's death is a disaster for everyone who dreams of free and peaceful Russia

Reposted by Dmitry Antipov

Sergey Nurk @sergeynurk.bsky.social · Feb 10

One more opening in my group at Oxford Nanopore. This time in our New York office!
ejnh.fa.em2.oraclecloud.com/hcmUI/Candid...

Dmitry Antipov @dantipov.bsky.social · Feb 8

Finally verkko 2.0 is out! github.com/marbl/verkko Waiting for your bug reports;)

My dream for 46 t2t scaffolds on the de novo human genome assemblies out of the box is still a dream, but for some samples we are really close.

Dmitry Antipov @dantipov.bsky.social · Feb 1

Congratulations!

Dmitry Antipov @dantipov.bsky.social · Oct 27

Ok, got your email, thank you! nice "practical" solution but still have questions

Dmitry Antipov @dantipov.bsky.social · Oct 27

Consider a DAG with edges having both weight and length. Problem: find a path between given vertices with highest _average_ weight (= total_w/total_l). Is it known to be NP-hard in general case? Looks similar to 0-1 knapsack, but I failed to construct polynomial reduction..

Reposted by Dmitry Antipov

Adam Phillippy @aphillippy.bsky.social · Sep 12

"Minmers are a generalization of minimizers that enable unbiased local jaccard estimation" aka MashMap3 with Bryce Kille, @erikg.bsky.social, and @treangen.bsky.social is now accepted in Bioinformatics! Many thanks to the reviewers for their very careful reviews doi.org/10.1093/bioi...

Dmitry Antipov @dantipov.bsky.social · Sep 8

Ok, let's give this one a try too