No, but it sure can train students to see unprompted Duo push notifications asking if they are logging in right now and hit "Yes" regardless.
23.02.2026 21:25
👍 8
🔁 0
💬 0
📌 0
No, but it sure can train students to see unprompted Duo push notifications asking if they are logging in right now and hit "Yes" regardless.
it has been [0] days since I've tried to emulate 64-bit atomic operations using 32-bit atomic ops in a system with relaxed memory ordering
Bit different from our usual work
A PhD applying metagenomics to identify microbial plastic degradation and metabolism, fully funded studentship (UK students only).
www.findaphd.com/phds/project...
Our latest vg release introduces GBZ v2 with better compression for sequences. I originally assumed that the total sequence length in a pangenome graph would be similar to the size of the genome. This does not hold in the full HPRC graphs due to unaligned centromeres.
Looking for a postdoc to build my new lab at TGen (Phoenix, AZ) focused on pangenome methods for cancer and complex disease. Full stack — from pangenome assembly and compression to association studies and somatic variant discovery. Reach out if interested! guarracinolab.github.io#join
Native Bio has made the @nytimes.com for the second time. The #IDsov movement is a tidal wave....and a solution for these difficult times. “We are at a fork in the road. Do not choose hate & do not choose unsustainability.” ~ 21st Century Maxim, Author Unknown www.nytimes.com/2026/02/05/s...
This sounds implausible. And it would result in the file existing but having no or the wrong content, right? I would not believe this robot.
This was such a fun project.
Adam's thread on it is a great summary!
Maybe I need to get Kimi k2.5 set up somehow so I can run it for a full day at reasonable cost, in a contained environment that can be allowed to turn evil and so doesn't constantly need me to approve things.
My Claude Code distracted me constantly when I tried to let it work while I was in a meeting, has me now more confused about the nature of the problem I was trying to get it to solve than when I started, and then went on strike for increased subscription fees after 2 hours.
Am I holding it wrong?
This is a chat box where you can try to yell at your paper until it is written.
Something that makes my LaTeX code snap together out of lose curly braces, or that hands me BibTeX keys based on which paper in my .bib smells most like the sentence I just wrote, might be useful.
This does not appear to be that thing.
Excellent piece about why race is not a biological category and how medical education needs to respond. www.bmj.com/content/392/...
In vg we are vendoring in htslib partly so we can make sure we use an htslib built with libdeflate support. For blocked-gzip at least you can stream the overall file just fine.
We've all heard of ramdisk, but have you met cachedisk?
Apply ASAP to be a "Genomics Senior Systems Architect" AKA Linux Herder at @ucscgenomics.bsky.social !
The first test is convincing www.ucsc.edu/careers/ to show you #job number 82880.
Are #redhat #quay.io Docker container builds down for anyone else? Mine are failing with no logs, and then my trigger gets "automatically disabled due to successive internal errors".
You find yourself in the middle of a fastq file. You see + and @ all over the place. How do you find the start of the current record?
If you wrap and let the line lengths be inconsistent, or if you wrap and are allowing the sequence ID on the separator line, I think you can't do it. You can pad out the separator line to full length or otherwise make it eligible to be a quality line, and you need to seek all the way to the start.
I'm only considering nice FASTQs that don't wrap or use IDs on the separator line, which I think are all that a lot of our tools support.
If you want to support wrapped FASTQ with consistent line lengths, I think you can still do it. Or if you want separator IDs. But not both.
Then, if not a quality or a separator, you're either sequence or a header. If your first character is "@", you're a header. Otherwise, you're sequence.
Then, if not a quality, try and eliminate being a separator. If your line is anything other than "+", you're not a separator. If it's "+" and you're not a quality, you're a separator.
I think you go by process of elimination.
First try to eliminate being in a quality string.
- Previous line must be "+"
- Line before that must be same length as your line, all bases
- Line before that must start with "@"
If any fails, you're not a quality. Otherwise, you are.
Well the line you're on is a header, sequence, a separator, or a quality string.
You can look at the character you're at and rule out some of those.
You can look at the next and previous characters and expand out, and eventually work out your line type and find the corresponding header. Probably.
But we call it the "Heng Effect" or "Heng Li Effect" all the time, just never in the literature.
Anyway, @lh3lh3.bsky.social pointed out that this could happen/was happening in an email once, and we spent like a year scratching our heads about it, and invented all sorts of frequency filtering approaches and (finally?) pangenome personalization (doi.org/10.1038/s415...) to try and deal with it.
Did anybody ever actually write up the "Heng Effect" (where adding variants to your pangenome graph can *reduce* its usefulness for samples unlikely to contain those variants)? Or is it just deep vg lore known only in our lab?
Kiddo age 5: daddy you don't work, you just sit at the computer
Kiddo age 12: dad you just write grants
SMBE2026 Symposium 21 | From genomic graphs to evolutionary insights: standardising pangenomes for population inference
SMBE2026 Symposium 21 | From genomic graphs to evolutionary insights: standardising pangenomes for population inference
📨 Abstract submission
smbe2026.org/abstracts
📋 Programme details
smbe2026.org/programme
#SMBE2026
The scikit-bio paper in online in Nature Methods! Many thanks to our collaborators, community contributors and reviewers! We couldn’t have done it without you. www.nature.com/articles/s41... #Bioinformatics #OpenSource