Good morning Bluesky ☀️ My grandad told me he saw the Titanic - from the outset he warned all the people that the ship would sink, but nobody listened to him. He knew he was right and didn't give up. He warned them again & again, until eventually they kicked him out of the cinema.
Nick Loman
@pathogenomenick.bsky.social
Public health (meta)genomics & bioinformatics, Prof @unibirmingham.bsky.social, Director @imibirmingham.bsky.social & @climb.ac.uk
People using AI to write: your readers can tell. Your texts become insufferably boring, full of AI-isms (e.g.: it's not this -- it's that) and of stupid statements that seem logical to a machine but are just dumb truisms for humans. Please, don't.
1st genome sequenced (Sanger 1977), 1st genome synthesised (Venter 2003), 1st genome + proteome fully mutated (Huijin Xiangua 2026!) @crg.eu @sangerinstitute.bsky.social Complete Mutagenesis of the Genome and Proteome of ΦX174 www.biorxiv.org/content/10.6...
Complete Mutagenesis of the Genome and Proteome of ΦX174
The bacteriophage ΦX174 was the first genome to be sequenced and the first to be chemically synthesised. Here we present a complete map of the consequences of changing every nucleotide in the ΦX174 ge...
biorxiv.org
"No amount of shorter showers or diligent recycling can substitute for decarbonising electricity, transport, industry and domestic heating." theconversation.com/was-this-hea...
‘Was this heatwave caused by climate change?’ – we’re asking the wrong question
Instead of asking whether climate change caused a heatwave, we should ask how much hotter it became because of fossil fuel emissions.
theconversation.com
It’s now official: June–July 2026 was the most anomalously warm two-month period ever recorded in France by a staggering margin. France is now virtually certain to record its warmest summer on record.
Very interesting paper on smallpox sequences from when they were ibtrocuced to South America, capturing the complex evolutionary forces at play in the pathogen city of this virus. www.science.org/doi/10.1126/...
The genomic identity of early smallpox in South America
Smallpox was a major driver of population collapse in the Americas after European contact, yet the genetic identity of the causal strains remains unknown. Here, we report the first ancient smallpox ge...
science.org
Geez, next thing you know we’ll be finding out the Reflecting Pool algae was a naturally occurring phenomenon and not a form of aquatic chemical warfare launched under cover of darkness by a team of highly trained, radical antifa frogmen.
The overwhelming right wing sentiment is that COVID was an engineered bioweapon that was also harmless. Trump overreacted to it by shutting down the country, then engineered a drive to create a vaccine that was the real bioweapon that killed millions, and it's Joe Biden's fault.
Microsoft-owned LinkedIn just added a “seems like AI slop” button to report posts that don’t feel human-written. It’s part of a broader push to get rid of AI slop on LinkedIn www.theverge.com/ai-artificia...
LinkedIn actually adds a ‘seems like AI slop’ button
It’s also removing an AI “enhance” feature.
theverge.com
Peter's Law of Chatbot Accuracy*: "everything you read that's produced by an LLM is absolutely true, except for the rare output about which you happen have expert knowledge". *Remixed from Knoll's Law of Media Accuracy, where the exact same rules apply!
AI is a normal tech. That means there are times when it is useful and times when it is pointless or actively harmful. One rule of thumb for chatbots is that they can only provide useful information to experts who have the discernment to ignore the defective output that LLMs *always* emit. 60/
Idea: during summer break we invert the system and have an ‘actually in the office auto reply’
I've been using Sanger's Artemis to view #genomics since ... who knows (it was first published in 2000). It's still super useful, but getting harder to install. I vibe-coded a lightweight browser-only version that views #GenBank files: linsalrob.github.io/genbank_view... 🧬💻 #phagesky
genbank_viewer
linsalrob.github.io
Genome-resolved metagenomics of traditional fermented beverages reveals biosynthetic diversity and informs the rational in silico design of probiotic synthetic communities https://www.biorxiv.org/content/10.64898/2026.07.22.740150v1
What is an 'everyday task' and what is a 'complex problem' in bioinformatics software engineering? Feels more important use the right model now they're not so heavily subsidised! I did a small comparison on an optimisation task: johnlees.me/posts/ska2-c... Small/cheap model wasn't good enough
Credit use of AI models for bioinformatics: a tiny benchmark
Agentic coding models such as codex and claude code are really useful for software engineering tasks. We sadly (and unsurprisingly) seem to be exiting the era of highly subsidised costs to get us hook...
johnlees.me
Genome Science UK 2026 nerve centre 🧬🧬 - Manchester share.google/aFa3SxwFvhtt... @pathogenomenick.bsky.social
This is so important - allowing individual uses of data to be tracked as citations will be so useful for demonstrating impact and reach of genome data. When you use data from Pathoplexus, create a SetSet & DOI for the sequences used and cite it in your paper.
2/ 📄 SeqSets/DOIs now show citations! We've integrated with @crossref.bsky.social so you can see which scientific publications have used a SeqSet - on the SeqSet page *and* on individual sequence pages. Using seqs in a paper? Don't forget to cite the DOI! 👉 pathoplexus.org/about/citation
"An insecure sandbox is not a demonstration of cyberprowess." -Loren Kohnfelder (on Openai's model "escaping containment" and attacking Hugging Face's servers) designingsecuresoftware.com/writings/com...
Normalizing cybersecurity facepalms
OpenAI writes: “Last week, Hugging Face disclosed a new kind of security incident(opens in a new window) after they detected and contained an AI agent that compromised their infrastructure, something...
designingsecuresoftware.com
Could Excel be conscious? We surveyed more than Jan 1 1900 people to find out.
woof ≈80% of assignments this summer term have essentially been just raw chatgpt/claude output, including assignments that required designing stuff in illustrator/affinity/canva. The battle is lost for asynchronous online courses.
New post, on why people shouldn't 'never-skill' themselves with AI: kucharski.substack.com/p/dont-never...
Don’t 'never-skill' yourself with AI
If you rent your abilities from AI companies, other people can rent them too
kucharski.substack.com
New strategy: ask prospective applicants to write their best idea for how to make a bioweapon in the application
I wish I could teach people just how grating it is to get the fourth clearly AI job application of the week that reads like "While I'm interested in X, to be honest what truly drew me to your lab was Y. You not only do A; you also do B."
Our paper out in Science today: The Metis defense system senses molecular byproducts released when a phage degrades the bacterial genome, and then prevents replication of the phage in the infected cell Congratulations @ostermanilya.bsky.social and co-authors!
Bacteria sense virus-induced genome degradation via methylated mononucleotides
Phages often degrade the genome of their bacterial host to individual nucleotides. Here we describe Metis, a bacterial defense system that directly senses phage-mediated host genome degradation. Metis...
science.org
Excited to share that our Metis story is now published in Science! 🎉 Bacteria can sense phage-induced degradation of their genome and activate immunity. Many thanks to everyone who contributed to this work. 📄 www.science.org/doi/10.1126/... More details in my earlier thread 🧵⬇️
Very proud of this preprint out of my lab, our first (of hopefully many) host-microbe spatial transcriptomics paper. www.biorxiv.org/content/10.6... Amazing insights into P. aeruginosa biology and chronic airway disease @cysticfibrosis.org.uk @sangerinstitute.bsky.social @asthmaandlung.org.uk
Important post on the ongoing Bundibugyo outbreak in DRC/Uganda. The clock appears to be fairly well-calibrated by now, even if a little slower than what we might have expected. Timing ~early-mid March and a bit of uncertainty re. 1 or possibly 2 intros. virological.org/t/genomic-ep...
Genomic epidemiology of the ongoing 2026 Bundibugyo Virus Disease outbreak in the Democratic Republic of the Congo
Genomic epidemiology of the ongoing 2026 Bundibugyo Virus Disease outbreak in the Democratic Republic of the Congo Institut National de Recherche Biomédicale (INRB), Kinshasa – DRC and partners Conte...
virological.org
Surprised LinkedIn so low….
As much as 40 percent of longform posts on LinkedIn and X that users actually see are fully AI-generated, according to new research Particularly notable because we know these platforms are flooded with slop, but this studied what the algorithm actually shows people www.404media.co/linkedin-and...
In some recent blog posts, I had a Klebsiella genome with a mysteriously higher ONT error rate. But Michael Biggel (University of Zurich) emailed me with the likely explanation: 7-deazaguanine mods. Thanks, Michael! Here's his preprint about this: www.biorxiv.org/content/10.6...
Standalone nanopore sequencing for foodborne pathogen surveillance: a large-scale evaluation and quality control framework
Whole-genome sequencing (WGS) is central to foodborne pathogen surveillance and cross-border outbreak detection. Long-read sequencing using Oxford Nanopore Technologies (ONT) promises rapid, complete,...
biorxiv.org
139 genomes now sequenced from the ongoing 2026 Bundibugyo virus disease outbreak in DRC, spanning 16 Health Zones. Led by INRB with support from the ARTIC network (Wellcome Trust and Gates Foundation). Full analysis: virological.org/t/genomic-ep...
Genomic epidemiology of the ongoing 2026 Bundibugyo Virus Disease outbreak in the Democratic Republic of the Congo
Genomic epidemiology of the ongoing 2026 Bundibugyo Virus Disease outbreak in the Democratic Republic of the Congo Institut National de Recherche Biomédicale (INRB), Kinshasa – DRC and partners Conte...
virological.org
Great reporting on this perspective in Science: www.science.org/doi/10.1126/... 1. Risks of ultra-processed foods are likely caused by their differences in specific nutritional properties, not in being 'processed' 2. 'Ultra-processed' miscategorizes many healthy + sustainable plant-based products
Ultraprocessed foods and obesity: Interpreting the evidence
What do randomized controlled trials really tell us about the effects of ultraprocessed foods?
science.org
New analysis suggests other factors may explain the harms blamed on “ultra-processing.” www.vox.com/future-perfe...
"a universal approach to quality assurance of research is no longer appropriate...we propose a layered approach...mov[ing] away from the norm all research must be peer-reviewed [to] a combination of trust markers, peer engagement, and evidence synthesis" researchonresearch.org/the-future-o...
The Future of Peer Review
A major new RoRI report reveals that peer review is at a breaking point and outlines an ambitious blueprint for system-wide reform
researchonresearch.org