Want to join a global consortium using #singlecell technologies to understand human diversity? Are #AI and #MachineLearning your #s? Come to Singapore! We're looking for leaders to drive v1 of the Human Diversity Atlas within the #HumanCellAtlas. tinyurl.com/5ccsfffz

Congrats, @Al__Forrest - this is a tour de force! Having a cleaned up database of ligand-receptor pairs is a great boost for signaling inference from omics data.

A 90° flip. A delayed hardening. Two tiny tweaks in hip growth millions of years ago helped our ancestors stand tall. sciencenews.org/article/gene…

Couldn’t agree more. It’s very useful to bin cellular diversity into discrete “cell types” - this helps us build mental models. But the reality is that cells exist on a phenotypic continuum.

All in all, sex-specific genetic analyses haven't led to any major discoveries yet. Despite massive differences in disease epidemiology by sex, effects of genetic variants on disease risk seem to be largely consistent between men and women.

Wow. 26% of MSS colorectal cancer patients showed some response to immunotherapy, and 20% had at least ten-fold reduction in tumor volume. Still Phase 2, but this could be transformative for CRC! Huge congrats to the team.

30 million American adults and 2 million adolescents and young adults (ages 12–25) take these medications, yet their adverse cardiometabolic effects often go unaddressed by healthcare providers.

A 42-country analysis of change in cancer patterns with age indicates only colon cancer is clearly on the rise in the young (age 20-49 years). Most of the rest of 12 cancer types were increasing across the age groups. AAPC-average annual percentage change acpjournals.org/doi/10.7326/…

Emirati Genome Program: - whole genome data from 504,000 individuals - electronic health record data from 426,000 - genetically reconstructed pedigrees applied here to understanding retinal diseases medrxiv.org/content/10.1101/…

I keep saying these guys are trouble.

A newly unearthed partial fossil skeleton suggests that an extinct hominid species could have made stone tools around 1.5 million years ago. sciencenews.org/article/foss…

In science, who gets grants and pubs depends on peer review. Does reviewer diversity matter? Yes! In new paper led by @jmzumeldumlao we establish a subtle structural bias we call geographical representation bias that benefits authors from wealthier countries (1/3)

🌟GLOW-ing together for better gut health A*STAR Genome Institute of Singapore (@astar_gis) is partnering clinicians and scientists from Institute of Mental Health, NHG Polyclinics, Tan Tock Seng Hospital, @dukenus, National University of Singapore, Lee Kong Chian School of Medicine, and PRECISE to explore how gut microbiomes in Asian populations may influence mental health. This multi-institutional study aims to uncover biological links between our gut and brain—potentially transforming how we understand and treat mental health conditions. 👉a-star.edu.sg/gis/news-event… Stay tuned as science dives deeper into the gut-brain connection! 🧬💡 #GutMicrobiome #MentalHealth #PrecisionMedicine #SingaporeResearch #ASTAR #GIS

Exciting work! Accurate risk prediction for common age-related diseases like CVD, stroke etc could eventually transform healthcare by enabling prevention and early detection.

Omics showed the power of profiling many genes side by side. Same logic applies here: characterizing multiple disease models in parallel gives more precise insights. Single cell comparative organoidomics of neurodevelopmental disorders is a very promising strategy!

We are super proud of our @YaleMed and Clinical-oMx Lab @UQ_News team for our new publication in @NatureGenet titled “Spatial signatures for predicting immunotherapy outcomes using multi-omics in non-small cell lung cancer”. Let’s get into it…nature.com/articles/s41588-0…

Years ago we tried to model cancer-associated fibroblast (CAF) subtypes in vitro, but gave up because only the proliferative subtype survived culture (duh!). This study could change all that: Perturb-seq with CRISPRa of 1,836 TFs induces diverse fibroblast states in vitro(!). 👏

We’re opening several postdoctoral and scientist positions focused on RNA structure prediction and mRNA drug optimisation at the @astar_gis. If you’re passionate about AI for therapeutics — especially in areas like generative models, diffusion models, flow matching, or joint embeddings — DM me. 🔗 Open positions: 🔹 AI Scientist, Generative AI Models for the Creation of RNA-Based Vaccines (GIS) - lnkd.in/gRA2mSiA 🔹 Scientist, AI in Drug Discovery (GIS) - lnkd.in/gkASi3Vi #AI #RNA #therapeutics

Finding somatic mutations in healthy tissue that can lead to cancer, at scale. Yet another path to primary prevention nature.com/articles/s41586-0… nature.com/articles/d41586-0…

What if your mitochondria could leave your cell… and rescue someone else’s? (Spoiler: they can. But it’s not always a rescue.) : This blew me away. A Nature paper just showed that brown fat cells, the ones you activate with cold exposure , can export mitochondria directly into the bloodstream. Whole mitochondria. Still breathing. And guess what? Other cells including damaged, energy-depleted ones can pull those mitochondria in and start working again. Like little organelle transfusions. You don’t just make energy. You can share it. But here’s the catch. Those same mitochondria can also land in tumor cells… and fuel their growth. Because mitochondria don’t have moral alignment they have voltage. That’s where it gets wild. From a field lens, this means: •Redox status matters •EZ water layers matter •UV/NIR signal balance matters •Even melatonin gradients might determine who gets the charge If the terrain is scrambled, those traveling mitochondria might turn from paramedics into accomplices. Easter Egg Hunt : Look up “mtDNA-deficient tumors” and how imported mitochondria restore OXPHOS. Then trace that back to circadian light, POMC signaling, and metabolic field coherence. Golden Nugget: Brown fat might be your field hospital, but your redox terrain decides whether it heals… or fuels. Very Interesting. Stay tuned sports fans!

How can spatial transcriptomics achieve whole-genome coverage while maintaining single-molecule resolution without requiring sequencing?@CellCellPress @YaleMed "Sequencing-free whole-genome spatial transcriptomics at single-molecule resolution" • RAEFISH addresses the limitation that existing spatial transcriptomics methods either offer genome-wide coverage with lower spatial resolution (spatial-capture/tagging technologies) or single-molecule resolution with pre-selected limited gene sets (image-based methods like multiplexed FISH targeting typically hundreds to ~3,000 genes), preventing unbiased hypothesis-free analyses at high resolution. • The authors developed RAEFISH (reverse-padlock amplicon-encoding fluorescence in situ hybridization), an image-based method using reversed padlock probes with invariant ends that enable cost-efficient synthesis, rolling circle amplification for signal generation, and multiplexed FISH-based decoding with a 94-choose-4 coding scheme where only 4% of targets are imaged per round to minimize signal overlap; they demonstrate profiling of 23,000 human genes or 22,000 mouse genes with detection of 3,749 RNA molecules from 1,287 genes per A549 cell, 864 molecules from 494 genes per mouse liver cell, and achieved 8.3% detection efficiency relative to MERFISH while showing 0.66 correlation with bulk RNA-seq. • Probe libraries were synthesized through oligo pool amplification allowing 2,000+ experiments from one $5,132 purchase ($158 per experiment vs. $19,407 for equivalent STARmap coverage), samples underwent reverse transcription, padlock ligation via splint oligo, rolling circle amplification, encoding probe hybridization, then 47 rounds of 2-color readout FISH imaging for 94-bit decoding using custom wide-field microscopes with 60× objectives capturing 107.9 nm pixels; cell segmentation used Cellpose/Cellpose3 for different tissue types, with ClusterMap for liver, and correlation analyses used Pearson coefficients within assays and Spearman between assays. • In A549 cells, RAEFISH identified 1,006 cell-cycle-associated genes including known markers (CKS2, CENPF, CDC20 for G2/M; RAD51, RRM1, TYMS for S phase) and 99 cell-cycle-associated lncRNAs, with lncRNAs showing higher nuclear ratios than protein-coding genes; in mouse liver (38,338 cells from 2 replicates), all major cell types were identified with clear periportal/pericentral zonation patterns, 65-208 zonation markers per cell type including cell-type-invariant markers (Cyp2f2, Cyp2e1, Glul) and cholangiocyte-specific markers (39 pericentral markers enriched for retinol/terpenoid metabolism), plus enriched cholangiocyte-leukocyte interactions with cholangiocytes expressing MHC class II molecules (H2-Eb1, H2-Aa, H2-Ab1, Cd74); in mouse placenta (34,582 cells), decidual macrophages showed higher MHC class II expression while Hofbauer cells expressed angiogenesis genes (Vegfa, Pdgfb), DSCs neighboring vascular ECs upregulated chemerin (Rarres2) and oxidative stress protection genes (Cryab, Txnrd1, Prdx5); in mouse lymph node (102,960 cells), clear B/T cell zone separation was observed with 2 major groups of zonation markers related to antigen presentation (positive markers) and T cell differentiation (negative markers); Perturb-RAEFISH screening 574 genes with 3 gRNAs each in A549 cells detected average 41 gRNA copies per cell with 89.3% single-gRNA dominance, validating target knockdown and identifying 552 perturbation genes significantly affecting expression of 492 MERFISH-probed genes, with hierarchical clustering revealing 52 perturbation clusters and 43 response clusters associated with specific biological processes (e.g., cluster 6: sister chromatid cohesion; cluster 8: DNA damage response). Authors: @ChengYubao , Shengyuan Dang, Yuan Zhang et. al @SStevenWang Link:cell.com/cell/fulltext/S0092…