Philipp Bayer
@philippbayer.bsky.social
Bioinformatics / tech / books
'It is required that one man should goof gloriously for the people.'
'It is required that one man should goof gloriously for the people.'
Reposted by Philipp Bayer
🎯 Marine sites wanted!
Join eDNA Expeditions, an OBIS-led project mobilizing environmental DNA to support local marine biodiversity monitoring and increase global ocean knowledge.
☑️ Kits, training & support provided
✅ Don't miss the opportunity!
➡️ ednaexpeditions.org
Join eDNA Expeditions, an OBIS-led project mobilizing environmental DNA to support local marine biodiversity monitoring and increase global ocean knowledge.
☑️ Kits, training & support provided
✅ Don't miss the opportunity!
➡️ ednaexpeditions.org
January 12, 2026 at 1:49 PM
🎯 Marine sites wanted!
Join eDNA Expeditions, an OBIS-led project mobilizing environmental DNA to support local marine biodiversity monitoring and increase global ocean knowledge.
☑️ Kits, training & support provided
✅ Don't miss the opportunity!
➡️ ednaexpeditions.org
Join eDNA Expeditions, an OBIS-led project mobilizing environmental DNA to support local marine biodiversity monitoring and increase global ocean knowledge.
☑️ Kits, training & support provided
✅ Don't miss the opportunity!
➡️ ednaexpeditions.org
Which country, outside of a China, has seen this as an opportunity to bolster their own scientific work-force? It looks like only China has taken up the opportunity (at scale) of becoming the new scientific leader. Other countries just crickets
This is the most astonishing graph of what the Trump regime has done to US science. They have destroyed the federal science workforce across the board. The negative impacts on Americans will be felt for generations, and the US might never be the same again.
www.nature.com/immersive/d4...
www.nature.com/immersive/d4...
January 20, 2026 at 11:32 PM
Which country, outside of a China, has seen this as an opportunity to bolster their own scientific work-force? It looks like only China has taken up the opportunity (at scale) of becoming the new scientific leader. Other countries just crickets
Reposted by Philipp Bayer
This is the most astonishing graph of what the Trump regime has done to US science. They have destroyed the federal science workforce across the board. The negative impacts on Americans will be felt for generations, and the US might never be the same again.
www.nature.com/immersive/d4...
www.nature.com/immersive/d4...
January 20, 2026 at 10:53 PM
This is the most astonishing graph of what the Trump regime has done to US science. They have destroyed the federal science workforce across the board. The negative impacts on Americans will be felt for generations, and the US might never be the same again.
www.nature.com/immersive/d4...
www.nature.com/immersive/d4...
Reposted by Philipp Bayer
My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring!
mirdita.org
mirdita.org
Mirdita Lab - Laboratory for Computational Biology & Molecular Machine Learning
Mirdita Lab builds scalable bioinformatics methods.
mirdita.org
January 20, 2026 at 11:07 AM
My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring!
mirdita.org
mirdita.org
Reposted by Philipp Bayer
1,000 PacBio genomes in a prospectively designed clinical utility study.
This was the biggest and most important study that made us go live in diagnostics.
Long-read genomes as a genetic first tier test across many rare diseases!
www.medrxiv.org/content/10.6...
This was the biggest and most important study that made us go live in diagnostics.
Long-read genomes as a genetic first tier test across many rare diseases!
www.medrxiv.org/content/10.6...
Clinical long-read genome sequencing for rare disease diagnostics
Background Diagnostic evaluation of rare genetic disorders continues to rely on multiple test modalities, despite the increasing use of short-read exome or genome sequencing as first-tier tests. Long-...
www.medrxiv.org
January 19, 2026 at 8:30 AM
1,000 PacBio genomes in a prospectively designed clinical utility study.
This was the biggest and most important study that made us go live in diagnostics.
Long-read genomes as a genetic first tier test across many rare diseases!
www.medrxiv.org/content/10.6...
This was the biggest and most important study that made us go live in diagnostics.
Long-read genomes as a genetic first tier test across many rare diseases!
www.medrxiv.org/content/10.6...
Spent the entire afternoon getting a company-internal Quarto theme to work. It's just a regular Quarto extension, loaded via the yaml. But Quarto render completely ignores the header. tried every possible thing. I think I'm going insane
January 19, 2026 at 10:46 AM
Spent the entire afternoon getting a company-internal Quarto theme to work. It's just a regular Quarto extension, loaded via the yaml. But Quarto render completely ignores the header. tried every possible thing. I think I'm going insane
It has finally happened: qs was removed from CRAN. I only realised because I tried to build my fancy-pants targets pipeline from start
(the solution is to use qs2, but it seems like targets does not yet support qs2?)
#rstats
(the solution is to use qs2, but it seems like targets does not yet support qs2?)
#rstats
January 19, 2026 at 7:41 AM
It has finally happened: qs was removed from CRAN. I only realised because I tried to build my fancy-pants targets pipeline from start
(the solution is to use qs2, but it seems like targets does not yet support qs2?)
#rstats
(the solution is to use qs2, but it seems like targets does not yet support qs2?)
#rstats
Reposted by Philipp Bayer
Weekend thoughts on Gas Town, Beads, slop AI browsers, and AI-generated PRs flooding overwhelmed maintainers. I don't think we're ready for our new powers we're wielding. lucumr.pocoo.org/2026/1/18/ag...
Agent Psychosis: Are We Going Insane?
What’s going on with the AI builder community right now?
lucumr.pocoo.org
January 18, 2026 at 10:38 AM
Weekend thoughts on Gas Town, Beads, slop AI browsers, and AI-generated PRs flooding overwhelmed maintainers. I don't think we're ready for our new powers we're wielding. lucumr.pocoo.org/2026/1/18/ag...
Reposted by Philipp Bayer
We mourn the passing of Peer Bork, EMBO Member since 2000: https://www.embl.org/news/embl-announcements/in-remembrance-of-peer-bork/
In remembrance of Peer Bork | EMBL
EMBL and its community are deeply saddened by the death of Peer Bork, the organisation’s Interim Director General.
www.embl.org
January 17, 2026 at 2:52 AM
We mourn the passing of Peer Bork, EMBO Member since 2000: https://www.embl.org/news/embl-announcements/in-remembrance-of-peer-bork/
Belated Christmas miracle in my house: the can of WD40 was where it was supposed to be! I thought by now the WD40 gremlins had whisked it away, never to be found again
January 18, 2026 at 7:05 AM
Belated Christmas miracle in my house: the can of WD40 was where it was supposed to be! I thought by now the WD40 gremlins had whisked it away, never to be found again
Found a weird way of working with ML:
- have LLM write image processing Python that calls on classic, local CNNs via preprocessing
- upload the preprocessed images and the boundary box images to the LLM
- have LLM adjust the Python for the CNN
- repeat until results good
- have LLM write image processing Python that calls on classic, local CNNs via preprocessing
- upload the preprocessed images and the boundary box images to the LLM
- have LLM adjust the Python for the CNN
- repeat until results good
January 15, 2026 at 12:22 PM
Found a weird way of working with ML:
- have LLM write image processing Python that calls on classic, local CNNs via preprocessing
- upload the preprocessed images and the boundary box images to the LLM
- have LLM adjust the Python for the CNN
- repeat until results good
- have LLM write image processing Python that calls on classic, local CNNs via preprocessing
- upload the preprocessed images and the boundary box images to the LLM
- have LLM adjust the Python for the CNN
- repeat until results good
This year will be my tenth anniversary since finishing my PhD. In that time I've worked in academia, then non-profit, now for-profit.... what's left? Government? Consulting on my own?
January 14, 2026 at 1:43 PM
This year will be my tenth anniversary since finishing my PhD. In that time I've worked in academia, then non-profit, now for-profit.... what's left? Government? Consulting on my own?
LLMs were invented so I don't have to learn awk
January 12, 2026 at 6:53 AM
LLMs were invented so I don't have to learn awk
Reposted by Philipp Bayer
⁉️The ARC has delayed outcomes of ALL grants 1–4 months & increased scheduled outcome windows from 2 weeks to 3 months!
This reverses 4 years of progress in providing greater certainty & ability to plan for researchers, their families & unis.
Their excuse? Security checks under new ARC legislation👇
This reverses 4 years of progress in providing greater certainty & ability to plan for researchers, their families & unis.
Their excuse? Security checks under new ARC legislation👇
January 12, 2026 at 1:17 AM
⁉️The ARC has delayed outcomes of ALL grants 1–4 months & increased scheduled outcome windows from 2 weeks to 3 months!
This reverses 4 years of progress in providing greater certainty & ability to plan for researchers, their families & unis.
Their excuse? Security checks under new ARC legislation👇
This reverses 4 years of progress in providing greater certainty & ability to plan for researchers, their families & unis.
Their excuse? Security checks under new ARC legislation👇
Reposted by Philipp Bayer
Individuals whose phenotype deviates from genetic expectation defined by common variation are enriched for rare damaging variants in genes that cause rare disease www.medrxiv.org/content/10.6...
January 11, 2026 at 2:23 PM
Individuals whose phenotype deviates from genetic expectation defined by common variation are enriched for rare damaging variants in genes that cause rare disease www.medrxiv.org/content/10.6...
"It is simply impossible not to see the reality of what is happening. Writing code is no longer needed for the most part. It is now a lot more interesting to understand what to do, and how to do it"
I think I agree with this - we humans come up with the how and why, the LLM implements
I think I agree with this - we humans come up with the how and why, the LLM implements
January 11, 2026 at 10:22 AM
"It is simply impossible not to see the reality of what is happening. Writing code is no longer needed for the most part. It is now a lot more interesting to understand what to do, and how to do it"
I think I agree with this - we humans come up with the how and why, the LLM implements
I think I agree with this - we humans come up with the how and why, the LLM implements
I was studying Kanji and was wondering whether every single Kanji has the reading kou/コウ??
So I made a figure. Turns out, yes, kou/コウ is the most common reading in the Joyo kanji! but not by much
Pictured are how often the top 15 most common onyomi readings appear in the Joyo kanji.
So I made a figure. Turns out, yes, kou/コウ is the most common reading in the Joyo kanji! but not by much
Pictured are how often the top 15 most common onyomi readings appear in the Joyo kanji.
January 10, 2026 at 10:01 AM
I was studying Kanji and was wondering whether every single Kanji has the reading kou/コウ??
So I made a figure. Turns out, yes, kou/コウ is the most common reading in the Joyo kanji! but not by much
Pictured are how often the top 15 most common onyomi readings appear in the Joyo kanji.
So I made a figure. Turns out, yes, kou/コウ is the most common reading in the Joyo kanji! but not by much
Pictured are how often the top 15 most common onyomi readings appear in the Joyo kanji.
A machine cannot undergo psychoanalysis because a machine has no psyche. Please do better
Researchers who put major AI models through four weeks of psychoanalysis got haunting responses, from "abuse" to "fears of failure"
go.nature.com/49pxLU3
go.nature.com/49pxLU3
AI models were given four weeks of therapy: the results shocked researchers
Chatbots put through psychotherapy report trauma and abuse. Authors say models are doing more than role play, but researchers are sceptical.
go.nature.com
January 9, 2026 at 11:07 AM
A machine cannot undergo psychoanalysis because a machine has no psyche. Please do better
Reposted by Philipp Bayer
Now published in Nature Biotechnology:
go.nature.com/44P7nSm
If you missed it, the TL;DR is in my April thread below
go.nature.com/44P7nSm
If you missed it, the TL;DR is in my April thread below
January 6, 2026 at 9:38 AM
Now published in Nature Biotechnology:
go.nature.com/44P7nSm
If you missed it, the TL;DR is in my April thread below
go.nature.com/44P7nSm
If you missed it, the TL;DR is in my April thread below
it's OK to bully robots
January 4, 2026 at 4:24 PM
it's OK to bully robots
Learned a wonderful word tonight: 皮肉, ひにく, hiniku.
The first kanji means "skin", the second one means "meat".
You'd think it's some kind of culinary term then, right?
nope it means "irony"
The first kanji means "skin", the second one means "meat".
You'd think it's some kind of culinary term then, right?
nope it means "irony"
January 4, 2026 at 4:23 PM
Learned a wonderful word tonight: 皮肉, ひにく, hiniku.
The first kanji means "skin", the second one means "meat".
You'd think it's some kind of culinary term then, right?
nope it means "irony"
The first kanji means "skin", the second one means "meat".
You'd think it's some kind of culinary term then, right?
nope it means "irony"
Going to the library for ❤️self-care❤️
January 4, 2026 at 4:44 AM
Going to the library for ❤️self-care❤️
Starting a new job tomorrow so now I have to choose a specific 'adult person' mask to wear and as we all know, a hard decision because I'll have to wear that one for years
January 4, 2026 at 1:53 AM
Starting a new job tomorrow so now I have to choose a specific 'adult person' mask to wear and as we all know, a hard decision because I'll have to wear that one for years
Just landed back in Australia! I wonder what I missed while I was offline!
Oh. Oh
Oh. Oh
January 3, 2026 at 2:47 PM
Just landed back in Australia! I wonder what I missed while I was offline!
Oh. Oh
Oh. Oh