James Ferguson

Using libs or writing more specific optimised code?

Absolutely. I made kuva to enable another tool. So I don't really have that problem 😅

Flamegraph* (I should go to bed, it's 2am 😅)

Didn't try it. Only heard about it the other day from a comment on Reddit. I wrote kuva from a pretty ignorant place tbh. Though I think that also worked in my favour

Dude I couldn't believe it either! I thought I was going mad but you can search on the flange graph and there it was

Did a benchmark of my tool with plotters and then with kuva...Plotters was taking 20% of the time. Kuva, doing THE SAME PLOTS, takes 0.1% of the time in the flamegraph.

Well, that was worth it

haha. I always love the "wait...hang on...no, i'm EXCITED AGAIN" cycles it goes through.

Featherweight long read alignment using partitioned reference indexes - Scientific Reports Scientific Reports - Featherweight long read alignment using partitioned reference indexes

Partitioned indexes. Hasindu did this during his PhD.

www.nature.com/articles/s41...

Did you ever see the work from Hasindu on minimap2 split indexes called "featherweight"? Was handy when doing alignments across lots of small "pi" like devices (we used rock64 boards)

Barcode 100K Specimens: In a Single Nanopore Run It is a global priority to better manage the biosphere, but action must be informed by comprehensive data on the abundance and distribution of species. The acquisition of such information is currentl...

"Consensus sequences based on two ONT reads showed a mean divergence of 1.37% from their SMRT counterpart. However, divergences dropped rapidly, averaging 0.03% when based on five reads; this value was well below our target (0.10%)." #paraphrased

doi.org/10.1111/1755...

Yea it looks a lot better

New preprint from
@novoalab.bsky.social !

Which tRNAs are used by ribosomes during translation?

We introduce tRIBO-seq, a nanopore method to sequence ribosome-associated tRNAs and track how the active tRNA pool changes across stress conditions.

www.biorxiv.org/content/10.6...

Thread 👇

kuva v0.1.3 released

- Mostly bug fixes
- Changed the default font on everything to be sans serif (DejaVu Sans)
- A few backend changes

- preparation for new features. legend control, some new plots, and improving the backend/scene for web/gui use

github.com/Psy-Fer/kuva...

Volcano Plot for fun

I shall gladly take some extra tokens. I have ideas...

A comprehensive assessment of tandem repeat genotyping methods for Nanopore long-read genomes https://www.biorxiv.org/content/10.64898/2026.02.28.708646v1

and it's rust btw ;)

give me generators pleeeease. Really helps clean up code and modularise pipelining data. i'd use them day 1 in my current tools

ahh, love a good session of tradcoding on my shiny new tool (not the plotting one, a bioinf tool that uses that lib...which is why i built it)
But maaan, it's pretty slow. But there is no way in hell an LLM is touching my baby. It's a mess, it's fast as all hell, and it works super well.

oh dear...it might be

yea could work. but I'm handing that over to the scverse people when i'm done

I am aware i sound insane 😅

Now that my side quest (making a plotting library from scratch) is done, I guess i'll go back to finish the tool I wrote it for, and insert my scatter and bar plots 🤣

docs[dot]rs had an issue the last few days where it's builder wasn't picking up new crates, but now it's fixed (yay!) and now i'm watching my new crate climb the queue to get it's first doc build

737...and dropping

it is now doing a bit better and working through everything properly

yea, it can do it, but it is NOT an easy task. I am currently in the weeds sorting through all the various filters and methods that have lots and lots of caveats and interconnected reasoning that claude just straight up got wrong the first 10 times it tried to do it. After some manual interventions

Rawsamble: Overlapping Raw Nanopore Signals using a Hash-based Seeding Mechanism AbstractMotivation. Raw nanopore signal analysis is a common approach in genomics to provide fast and resource-efficient analysis without translating the s

This looks really interesting and perhaps might enable faster overlapping for HERRO @msikic.bsky.social , not sure academic.oup.com/bioinformati... , also not sure Rawsamble can overlap repetitive sequences, for which minimap2 can, but seems to slow down a lot compared to default settings

They look really pretty. I need to add step/lollypop plots.

So you can make up multiple heatmap plots. then set up your multifigure plot, and feed it a vector of thos plots and you have your easy multifigure plot
I need to doc that stuff but it's in the dos[.]rs docs..but there is an issue with the build atm cause they lost connection to crates[.]io

So here it is showing the labels in the boxes

and you use these for the axis lables
Layout::with_x_categories(labels Column labels on the x-axis
Layout::with_y_categories(labels) Row labels on the y-axis

yep the heatmap plot can do that.