iiinova — where your digital self keeps evolving

Vikram Shivakumar

@vikramshivakumar.bsky.social

180 followers 142 following 41 posts

PhD Student @ JHU Langmead Lab

Vikram Shivakumar @vikramshivakumar.bsky.social · 15/07/2026

6/ Shredtools extract depends on the coverage of multi-MUMs. Another method, enhance, finds additional matches between MUMs that serve as additional markers of the coordinate system, approaching a full multiple alignment. Figure: extract across the A. thaliana pangenome ("enhanced" MUMs in purple)

130

Vikram Shivakumar @vikramshivakumar.bsky.social · 15/07/2026

5/ are multi-MUMs useful as a coordinate system? We found that multi-MUMs are the ideal choice over unique k-mers, which have lower coverage. And often, the choice of k varies greatly with various parameters of the pangenome.

120

Vikram Shivakumar @vikramshivakumar.bsky.social · 15/07/2026

4/ Shredtools reports "bounds", the distance between the ouput region to the nearest flanking MUMs. Querying gene regions across HPRC2 assemblies, more than 80% of regions were immediately flanked by a multi-MUM, yielding exact coordinates across all 476 human assemblies!

110

Vikram Shivakumar @vikramshivakumar.bsky.social · 15/07/2026

2/ ⭐ highlight of the release: we implemented shredtools for the HPRC human pangenome as a browser-based, client-side app to query and fetch syntenic regions across the pangenome. You can even visualize synteny right in the browser! Read on for how the method works.

110

Vikram Shivakumar @vikramshivakumar.bsky.social · 21/07/2025

And of course, the poster itself:

030

Vikram Shivakumar @vikramshivakumar.bsky.social · 27/05/2025

We can also merge along the shape of a phylogenetic tree, finding clade-specific variation and conserved elements. Previously, adding new assemblies can lose MUMs, which must be present across the whole collection. Now we can find MUMs that reveal local variation distinct to specific subgroups. 3/n

142

Vikram Shivakumar @vikramshivakumar.bsky.social · 27/05/2025

We implement two partition/merge algorithms that can merge multi-MUMs between datasets. This makes Mumemto highly parallelizable, but also very memory efficient if partitions are computed in serial. 2/n

120

Vikram Shivakumar @vikramshivakumar.bsky.social · 09/05/2025

Excited to share our latest work on comparing and visualizing multiple genome assemblies to identify conservation and structural variation in pangenomes with Mumemto! Check out poster 250 at #bog25 if you are here. New preprint coming very soon 👀

03414

Vikram Shivakumar @vikramshivakumar.bsky.social · 26/02/2025

We ran Mumemto on 474 human assemblies from @humanpangenome.bsky.social to find syntenic regions using MUMs. Mumemto scales remarkably well to large pangenomes thanks to compressed-space algos! It took under 2 days across 7 nodes (each using ~500 GB memory).

195

Vikram Shivakumar @vikramshivakumar.bsky.social · 06/01/2025

p.p.s here's a logo I designed for Mumemto (pronounced like memento) inspired by the film.

130

Vikram Shivakumar @vikramshivakumar.bsky.social · 06/01/2025

Mumemto can visualize multi-MUM synteny to reveal large-scale pangenome structure. We also showed that multi-MUMs can reveal potential misassemblies (e.g one reported in HPRC v1) and scaffolding errors (contigs are oriented to match a reference w/o the pangenome context revealing a common inversion)

162

Vikram Shivakumar @vikramshivakumar.bsky.social · 06/01/2025

Mumemto uses prefix-free parsing, a compressed-space method to compute the enhance suffix array, to efficiently index large pangenome collections. This makes it quick and memory efficient, with remarkable scaling, enabling multi-MUM computation across hundreds of HPRC assemblies.

110