🧬 CRISPR just learned a new trick (and it's actually safer than the old one)

So we used to think that CRISPR could only edit DNA permanently, raising all sorts of "what if we mess this up" concerns? Turns out scientists just needed to delete 55 amino acids from an ancient protein to unlock a whole new superpower: editing RNA instead. This week we're covering the molecular Swiss Army knife that could fix genetic diseases without permanent changes, a blood test for chronic fatigue that has experts cautiously optimistic (and slightly skeptical), retinal imaging so precise it spots oxygen starvation in capillaries, a map of the research bottlenecks nobody will fund, and proof that science illustration shouldn't cost your entire grant budget.

RESEARCH

CRISPR proteins like Cas9 have spent years as DNA editors, which is great until you realize editing DNA is permanent, heritable, and carries the "oops, we changed the wrong thing forever" risk. Researchers at Yale just figured out how to reprogram these proteins to target RNA instead by deleting a 55-amino acid stretch from IscB, Cas9's evolutionary ancestor. Remove those amino acids (residues 433-487 if you're keeping score), and suddenly the protein loses its DNA obsession and starts cutting messenger RNA.

The breakthrough matters because RNA editing is inherently reversible. Mess up? The cell makes new mRNA anyway. No permanent genome changes, no heritable mutations, no existential dread about accidentally creating a genetic disaster for future generations. The truncated protein, called R-IscB, can do four things: splice RNA, correct mutations through trans-splicing, perform A-to-I base editing when fused to ADAR2, and destroy target mRNAs entirely. It binds RNA with sub-32 nM affinity and causes zero detectable cytotoxicity in human cells, which is a massive improvement over previous RNA editors that kept killing the cells they were supposed to help.

Prof. Ailong Ke's team at Yale discovered the secret was removing the TAM Interaction Domain, the molecular gatekeeper that normally recognizes DNA sequences. Without it, the protein's intrinsic affinity for single-stranded nucleic acids takes over. They tested the concept on four different Cas9 variants, and it worked every time, suggesting most DNA editors have been closet RNA editors all along.

We can already imagine the application! Correcting point mutations, fixing premature stop codons, silencing disease-causing genes, but without the permanence of DNA editing. The team successfully restored GFP fluorescence in human lung cells by fixing stop mutations. At one-third the size of typical Cas9 proteins, R-IscB fits easily into delivery vectors. First author Chengtao Xu notes working with IscB's simpler structure made the hidden potential obvious, which probably wouldn't have happened if they'd started with Cas9's "way more sophisticated" architecture. Sometimes evolution's rough draft is easier to reprogram than the polished version.

RESEARCH

Researchers at the University of East Anglia just published a blood test achieving 96% overall accuracy for ME/CFS by detecting 200 chromosome conformation biomarkers in blood cells. The EpiSwitch® 3D genomics test correctly identified 22 of 24 ME/CFS patients (92% sensitivity) and 44 of 45 healthy controls (98% specificity), which sounds revolutionary for a condition that currently takes 1-5+ years to diagnose through the medical equivalent of "we've ruled out everything else, so I guess it's chronic fatigue."

The test measures how DNA folds in three dimensions during illness, revealing dysregulation in interleukin-2, TNFα, neuroinflammatory pathways, toll-like receptor signaling, and JAK/STAT pathways. For 17-24 million people worldwide with ME/CFS, objective laboratory proof would be life-changing. One-quarter of patients are bedridden, and delayed diagnosis correlates with worse outcomes since early rest during acute illness improves recovery.

But independent experts are skeptical. Prof. Chris Ponting at Edinburgh points out that the study compared severely affected, housebound patients against healthy controls with good exercise tolerance, which is an easy distinction. The test hasn't been validated against patients with depression, fibromyalgia, or MS who present with similar symptoms. Prof. Kevin McConway emphasizes this is "a proof-of-concept study, not a working diagnostic tool, not yet." Statistical uncertainty means that 92% sensitivity could plausibly range from 73% to 99%.

Dr. Alastair Miller notes the study lacked appropriate disease controls and worries it's "yet another false dawn." Oxford BioDynamics funded the research and employs most authors, though the EpiSwitch platform has a proven track record in ALS, rheumatoid arthritis, and prostate cancer diagnostics. The company estimates the test would cost around £1,000. So: promising? Yes. Ready for clinical use? Not remotely. Revolutionary? Ask again after independent validation with proper control groups.

RESEARCH

Johns Hopkins and University of Pennsylvania researchers achieved what ophthalmology has wanted for decades: measuring oxygen levels in individual retinal capillaries while simultaneously capturing 3D structural images. Published in Neurophotonics, the dual-imaging system combines visible light optical coherence tomography with phosphorescence lifetime imaging to create depth-resolved oxygen maps showing exactly which microscopic blood vessels are oxygen-starved. This matters because the retina consumes oxygen at one of the highest rates of any tissue, and disruptions drive diabetic retinopathy, glaucoma, and age-related macular degeneration.

Previous methods could only measure oxygen in larger vessels near the optic nerve head, missing the capillary-level changes where early disease manifests. This system achieves capillary resolution in ~10 μm vessels with 1.3 μm depth resolution in water, far exceeding traditional near-infrared OCT. The VIS-OCT channel provides structural anatomy and blood flow, while PLIM-SLO directly measures oxygen partial pressure using the Oxyphor 2P phosphorescent probe developed at UPenn. Higher oxygen means faster phosphorescence decay; lower oxygen means slower decay.

Lead author Stephanie Nolen and Prof. Ji Yi demonstrated physiological validation: arterioles had the highest oxygen, venules the lowest, capillaries intermediate, exactly as expected. When inhaled oxygen varied (69%, 78%, 88% systemic saturation), retinal oxygen levels tracked predictably. The system captured acquisition in under 60 seconds and is non-destructive, allowing repeated measurements to track disease progression or treatment response.

Clinical applications are immediate for research while human translation proceeds. Mouse models of retinal disease can now reveal exactly how microvascular dysfunction contributes to vision loss. The team notes that with adaptive optics integration, the platform could achieve cellular resolution for improved diagnostics in human eye disease. Current retinal imaging is a multi-billion-dollar market, but no existing technology combines depth resolution, capillary-level precision, and direct oxygen measurement. The work was supported by four NIH grants (and an additional one from NSF), suggesting strong federal confidence in translational potential, which is reassuring given that most cool imaging tech stays in research hell forever.

NEWS

Convergent Research released their Fundamental Development Gap Map, cataloging approximately 100 scientific bottlenecks and several hundred foundational capabilities that need mid-scale coordinated projects to build, but don't fit traditional funding mechanisms. The interactive database at gap-map.org represents systematic identification of "neither profitable nor publishable" infrastructure: academia can't pursue it (low publication potential) and startups can't monetize it (no immediate revenue path), yet entire fields stall without these tools. The solution is Focused Research Organizations (FROs), finite-duration teams of 10-30 scientists operating like deep-tech startups but pursuing public goods, funded at $20-50M per project over 5-7 years.

Since 2021, ten FROs have launched with $50M+ backing from Eric and Wendy Schmidt, plus Ken Griffin. E11 Bio is mapping complete neural circuits across mammalian brains. Cultivarium released software predicting microbial growth conditions from genomes, expanding synthetic biology beyond model organisms. PTI is scaling protein analysis 100-fold. And the Lean FRO revolutionized mathematical theorem proving, which Google DeepMind used to solve International Math Olympiad problems at a silver medalist level.

The funding gap exists because institutional incentives are broken. Nessan Bermingham of Khosla Ventures notes about biotech: "The venture community and pharma and biotech cannot step in to fill this gap" left by NIH budget constraints. Venture capital needs applications "just a few years away from making money," while academic incentives revolve around individual publications, disfavoring large coordinated teams. Co-founder Adam Marblestone, formerly at MIT, conceived FROs while frustrated that "interdisciplinary, team science projects were being systematically overlooked."

The UK's ARIA (Advanced Research and Invention Agency) is the first governmental program funding FROs, planning 1-2 UK FROs in early 2026. The Federation of American Scientists is pushing US agencies to pilot FRO funding. With NIH facing budget cuts as the 21st Century Cures Act expires in 2026, alternative mechanisms are increasingly urgent. The thesis: "A few billion dollars in input could unlock trillions in societal value" by removing catalytic bottlenecks. Think Human Genome Project scale of impact, but at $20-50M instead of billions, and distributed across 100 opportunities. Of course, actually funding these remains optional.

RESOURCE

Arcadia Science published a library of 71 organism illustrations under the CC0 public domain license, giving researchers zero-restriction access to professional-quality scientific visuals spanning model organisms (mice, zebrafish, C. elegans, Drosophila) to emerging research species (tardigrades, amoebas, diverse algae). Each organism comes in three formats: basic silhouette, detailed tricolor with stroke, and textured raster with hand-drawn aesthetic, all with transparent backgrounds and scalable vector files where appropriate.

Created by Audrey Bell, Arcadia's in-house illustrator with an MSc with Distinction in Medical Art from the University of Dundee and a 2018 Gold Award from the Institute of Medical Illustrators, the library originated from their Zoogle project matching organisms to research problems. The CC0 license is more permissive than alternatives: no attribution required (though encouraged), commercial use allowed, modifications permitted, redistribution unrestricted. This contrasts with PhyloPic (8,000+ silhouettes but mixed licenses) and SciDraw (CC-BY 4.0 requiring attribution).

The practical impact addresses real resource gaps. Professional organism illustrations typically cost hundreds to thousands of dollars to commission, pricing out many researchers, educators, and science communicators. Few free resources cover non-model organisms despite a growing push to expand research beyond traditional species. Bell's background includes medical illustration for the Royal College of Surgeons, Oakland University William Beaumont School of Medicine, and Microsoft HoloLens projects.

Community response includes specific requests (Xenopus laevis frogs and axolotls topped the wishlist), and the embedded Airtable means real-time additions. Version 2 dropped on October 1, just six days after launch, demonstrating active development. The library is small compared to PhyloPic's decade of accumulation, but its multi-format approach, professional quality, and maximum-permissive licensing carve a distinct niche. For Arcadia, a for-profit biotech company, freely sharing internal resources demonstrates that open science and commercial viability can coexist. Though one suspects the gesture also functions as excellent marketing for a company trying to differentiate itself in a crowded space.

So there you have it, once again. CRISPR learned to edit RNA without killing cells, chronic fatigue got a blood test that might actually work (pending independent validation with proper controls), your eyeballs can now snitch on oxygen-starved capillaries before you notice vision problems, the research funding system has documented all the important work it refuses to fund, and science illustrations are finally free for people who don't have illustration budgets. Just a normal week in biotech, where progress happens despite the incentive structures.

Did the RNA editing breakthrough make you reconsider permanent genome modifications? Skeptical about the ME/CFS test or cautiously optimistic? Think Convergent Research's gap map will actually get funded, or is it just another well-intentioned document the system will ignore? Planning to download some free tardigrade illustrations for your next presentation?

Hit reply and let us know. We read every email while contemplating whether RNA editing will finally make gene therapy safe enough for mainstream use, or if we're just creating more ways to mess with biology.

Keep questioning everything (and maybe download those organism illustrations), Prateek & Jere

P.S. To our readers still using permanent genome edits: We see you. But also, RNA editing exists now. Maybe give your future descendants a break from your experimental phase?