Igor Martayan 🦋

The code is optimized for multiple architectures (including ARM) and is already available as a crate! You can try it at github.com/rust-seq/simd-…

We also paid special attention to the computation of canonical minimizers, to ensure that a sequence and its reverse-complement always select the same minimizer positions. This constraint adds a 50% overhead to the running time, resulting in 6s for a human genome.

I'm glad to announce that the simd-minimizers library is out! @curious_coding and I have been optimizing the computation of minimizers down to the smallest detail. The result is an order of magnitude faster than existing methods ; processing an entire human genome takes only 4s!

I am super excited to share our latest preprint with @yhhshby, @IgorMartayan, and Lucas Robidou on k-mer representations, introducing Hyperkmer and a novel k-mer counter, KFC! biorxiv.org/content/10.110… 🧵 A thread:

The next symposium of experimental algorithms (SEA'25) is going to be in Venice 😎 Amazing invited speakers: Sebastiano Vigna, Daniel Lemire, and Giulia Bernardini! + workshop on compressed self-indexes, featuring Roberto Grossi and Giovanni Manzini! regindex.github.io/sea2025.github…

The open-closed mod-minimizer algorithm biorxiv.org/cgi/content/sh… #biorxiv_bioinfo

The open-closed mod-minimizer is officially out, with @daniel_c0deb0t and @giulio_pibiri :D A single simple & practical & efficient scheme that is best in class for k>w (closing ~half the remaining gap to optimal), and close to the best there currently is for k<w. 1/5

This new paper is a follow up of the WABI'24 paper featuring the mod-minimizer: it introduces the "open-closed" mod-minimizer scheme. Fantastic collaboration with @curious_coding and @daniel_c0deb0t! (A bioRxiv preprint is forthcoming.) jermp.github.io/assets/pdf/pap… 1/16

Features: '2+o(1)' bits per canon. k-mer for nice k-mer sets (the spectrum like property) and then linear w/the superstr. len (and no assumptions on the k-mer set structure), exact membership queries in O(k) (single) / O(1) streaming, and linear-time construction.

FroM Superstring to Indexing: a space-efficient index for unconstrained k-mer sets using the Masked Burrows-Wheeler Transform (MBWT) biorxiv.org/cgi/content/sh… #biorxiv_bioinfo

Excited to share our new method GreedyMini+ to generate DNA minimizers with low expected or particular density for practical k & w values: biorxiv.org/content/10.110…. Joint work with my student Ido Tziony and my colleagues @shay_gol, @matanHakrausi, and Arseny Shur #minimizers

Results are pretty amazing! (M is classic mini; C is closed-syncmer, O is open syncmer, and OC is open+closed syncmer; the versions with the mod- are their "lifted" versions given by the mod-sampling framework.) 8/10

Minimizers' folks: some more news on the development of new sampling schemes. I and @curious_coding are very hyped about it! 🧵 1/10

We had the mod-minimizer (blue) and very soon the lower bound preprint will be up (lower red function). Iterating a bit on @daniel_c0deb0t's idea (thanks PyO3 😆), I found this new scheme (dotted red), which halves the gap towards the lower bound, and may well be near-optimal.

Building upon @curious_coding and @BKille's recent breakthrough in lower bounding the density of local forward selection schemes, my colleagues Golan and Kraus et al. further tightened the lower bound: sites.google.com/view/shaygolan…. What an exciting time to study selection schemes!

Allright, I think my blogpost on computing random minimizers is ready to share. Runs in just under 0.5s for a full human genome :) That's 60x-160x faster than some other implementations. As part of this, there's also a 30x faster ntHash implementation. curiouscoding.nl/posts/fast-min…

@shaomingfu @pashadag The program looks amazing! Any chance this will be recorded, or at least the slides shared?

Check out the incredible lineup of talks at WEMSA 2024: ccbb.psu.edu/workshop2024/. There is still time to submit posters by 7/20/2024 via forms.gle/S1D5Uy1rLmfwB4….

It's official: DSB '25 is going to be in Pisa, 5th and 6th of March! Can't wait for this. #Bioinformatics