Zebrafish heart repair uses distinct cardiomyocyte layers for growth and regeneration
On September 8, biologists described in eLife a thin layer of cardiomyocytes in zebrafish hearts that helps assemble the muscle wall and coronary vessels during growth.
After injury, other cardiomyocytes rebuild lost tissue; zebrafish hearts can regenerate muscle after ventricular tip removal; surviving cardiomyocytes change state, divide, and replace the lost tissue; the heart consists of several such layers.
Researchers mapped 1,668 cardiomyocytes of adult heart and found a group with active phlda2 gene; these cells lie in a thin layer between outer compact muscle and inner trabeculae.
They created a fish line where phlda2-active cells produce bacterial NTR enzyme and become sensitive to metronidazole; treatment removed about 96.9% of these cells from the ventricle.
In young fish, trabecular muscle area became roughly 54.7% smaller, outer compact layer lost proper organization, and coronary vessels fragmented; after 90 days the vascular pattern remained disrupted.
This layer participates in how the growing heart assembles muscle and vascular network.
In adult fish, ventricular tip was removed six days after treatment start; one week later the wound contained as many gata4-active cells (a marker of dividing cardiomyocytes during repair) as in controls.
By day 30 new muscle covered the injury; vessel density and scar area matched controls; the phlda2‑active layer did not recover even after 60 days, and its cells did not overlap with those activating gata4 during regeneration, indicating that heart growth and adult tissue repair rely on different cardiomyocyte populations.
🔗 Read original →
On September 8, biologists described in eLife a thin layer of cardiomyocytes in zebrafish hearts that helps assemble the muscle wall and coronary vessels during growth.
After injury, other cardiomyocytes rebuild lost tissue; zebrafish hearts can regenerate muscle after ventricular tip removal; surviving cardiomyocytes change state, divide, and replace the lost tissue; the heart consists of several such layers.
Researchers mapped 1,668 cardiomyocytes of adult heart and found a group with active phlda2 gene; these cells lie in a thin layer between outer compact muscle and inner trabeculae.
They created a fish line where phlda2-active cells produce bacterial NTR enzyme and become sensitive to metronidazole; treatment removed about 96.9% of these cells from the ventricle.
In young fish, trabecular muscle area became roughly 54.7% smaller, outer compact layer lost proper organization, and coronary vessels fragmented; after 90 days the vascular pattern remained disrupted.
This layer participates in how the growing heart assembles muscle and vascular network.
In adult fish, ventricular tip was removed six days after treatment start; one week later the wound contained as many gata4-active cells (a marker of dividing cardiomyocytes during repair) as in controls.
By day 30 new muscle covered the injury; vessel density and scar area matched controls; the phlda2‑active layer did not recover even after 60 days, and its cells did not overlap with those activating gata4 during regeneration, indicating that heart growth and adult tissue repair rely on different cardiomyocyte populations.
🔗 Read original →
Nature
Tbx5a lineage tracing shows cardiomyocyte plasticity during zebrafish heart regeneration
Nature Communications - It is not clear if it is the embryonic origin or anatomical location of cardiomyocytes that restrict their contribution to zebrafish heart regeneration. Here, the authors...
BioPharma APAC Maps Four Longevity Funding Channels in Asia‑Pacific
On September 9, BioPharma APAC released a map of four channels of longevity financing in the Asia‑Pacific region. The map separates funds that pay for therapy development, diagnostics, clinic services, and scientific projects linked to the region but financed or incorporated outside it. Venture‑tracker records show when an investor receives equity after a deal.
Payments from pharmaceutical companies for joint development, research grants, and clinic revenue from examinations go to different recipients and support distinct work, which the author keeps separate. Therapeutic programs in the map draw venture money, public capital, grants, and pharma partnership payments. Biological‑age diagnostics — tests that assess age‑related changes in the body — form a separate line.
Two additional lines cover clinics and services funded
🔗 Read original →
On September 9, BioPharma APAC released a map of four channels of longevity financing in the Asia‑Pacific region. The map separates funds that pay for therapy development, diagnostics, clinic services, and scientific projects linked to the region but financed or incorporated outside it. Venture‑tracker records show when an investor receives equity after a deal.
Payments from pharmaceutical companies for joint development, research grants, and clinic revenue from examinations go to different recipients and support distinct work, which the author keeps separate. Therapeutic programs in the map draw venture money, public capital, grants, and pharma partnership payments. Biological‑age diagnostics — tests that assess age‑related changes in the body — form a separate line.
Two additional lines cover clinics and services funded
🔗 Read original →
Nature
Recommendations for biomarker data collection in clinical trials by longevity biotechnology companies
npj Aging - Recommendations for biomarker data collection in clinical trials by longevity biotechnology companies
Radical Numerics launches Omnii for personalized mRNA cancer vaccine target selection
On September 8, Radical Numerics unveiled Omnii, a post‑trained model designed to pick targets for personalized mRNA cancer vaccines. The system takes a tumor’s mutational profile and the patient’s MHC variants, then selects vaccine targets and builds an mRNA cassette sequence.
Omnii first predicts whether a mutant peptide will be presented by the patient’s MHC class I molecules on the tumor cell surface. It then estimates the peptide’s immunogenicity, i.e., whether a T‑cell will recognize the presented fragment and mount a response. Only peptides that pass both presentation and immunogenicity checks can be included in the vaccine cassette.
In head‑to‑head comparisons, Omnii achieved an AUROC 0.939 for MHC class I presentation, outperforming BigMHC‑EL’s 0.933. For immunogenicity prediction, the model scored AUROC 0.750, compared with 0.558 for BigMHC‑IM.
In a simulation where 100 candidate peptides were evaluated, only 6% were truly immunogenic and the cassette had five slots. Omnii selected an average of 1.3 immunogenic targets per cassette, while BigMHC‑IM managed just 0.7.
After target selection, Omnii outputs the mRNA sequence encoding the chosen neoantigens. The model is currently available in research preview, and Radical Numerics is seeking partners in cancer immunology and vaccine development.
🔗 Read original →
On September 8, Radical Numerics unveiled Omnii, a post‑trained model designed to pick targets for personalized mRNA cancer vaccines. The system takes a tumor’s mutational profile and the patient’s MHC variants, then selects vaccine targets and builds an mRNA cassette sequence.
Omnii first predicts whether a mutant peptide will be presented by the patient’s MHC class I molecules on the tumor cell surface. It then estimates the peptide’s immunogenicity, i.e., whether a T‑cell will recognize the presented fragment and mount a response. Only peptides that pass both presentation and immunogenicity checks can be included in the vaccine cassette.
In head‑to‑head comparisons, Omnii achieved an AUROC 0.939 for MHC class I presentation, outperforming BigMHC‑EL’s 0.933. For immunogenicity prediction, the model scored AUROC 0.750, compared with 0.558 for BigMHC‑IM.
In a simulation where 100 candidate peptides were evaluated, only 6% were truly immunogenic and the cassette had five slots. Omnii selected an average of 1.3 immunogenic targets per cassette, while BigMHC‑IM managed just 0.7.
After target selection, Omnii outputs the mRNA sequence encoding the chosen neoantigens. The model is currently available in research preview, and Radical Numerics is seeking partners in cancer immunology and vaccine development.
🔗 Read original →
PubMed Central (PMC)
Artificial intelligence in peptide cancer vaccine design: from neoantigen discovery to immunogenicity prediction
Peptide-based cancer vaccines represent a promising immunotherapeutic strategy aimed at inducing tumor-specific immune responses through the targeting of tumor-associated antigens and neoantigens. Recent advances in next-generation sequencing and ...
IRIS RNA model predicts developmental signals to improve stem cell differentiation
The IRIS model uses RNA profiles to infer which developmental signals acted on a cell. It was described in Nature Methods, September 8 after being trained on human stem cells exposed to known combinations of six developmental signals and then tested on single‑cell data from mouse embryos.
An RNA profile records which genes are currently active, giving a snapshot of the cell’s state. A benchmark of 400 models on cell atlases showed that merely expanding collections of RNA profiles quickly ceases to improve predictive power.
IRIS first learns from cells with known signal combinations, then searches the RNA profile for the trace of a signaling pathway—the chain through which a cell responds to an external cue. Researchers cultured human embryonic stem cells, applied defined signal mixes at different stages, and measured single‑cell RNA after each condition to teach IRIS to recognize these traces.
When applied to mouse embryo data, IRIS recovered the expected order of signal combinations in lineages that give rise to the foregut and heart muscle. It then predicted heightened activity of WNT and BMP in the prospective respiratory mesenchyme surrounding the lung bud.
In mouse foregut culture, activating WNT expanded the domain where the respiratory mesenchymal markers Tbx4 and Foxf1 are expressed, while a WNT inhibitor erased Tbx4 expression. Guided by this finding, the authors added WNT stimulation on the fourth day of differentiation of human stem cells and maintained it, which raised TBX4 levels and kept intermediate WNT activity to preserve FOXF1.
🔗 Read original →
The IRIS model uses RNA profiles to infer which developmental signals acted on a cell. It was described in Nature Methods, September 8 after being trained on human stem cells exposed to known combinations of six developmental signals and then tested on single‑cell data from mouse embryos.
An RNA profile records which genes are currently active, giving a snapshot of the cell’s state. A benchmark of 400 models on cell atlases showed that merely expanding collections of RNA profiles quickly ceases to improve predictive power.
IRIS first learns from cells with known signal combinations, then searches the RNA profile for the trace of a signaling pathway—the chain through which a cell responds to an external cue. Researchers cultured human embryonic stem cells, applied defined signal mixes at different stages, and measured single‑cell RNA after each condition to teach IRIS to recognize these traces.
When applied to mouse embryo data, IRIS recovered the expected order of signal combinations in lineages that give rise to the foregut and heart muscle. It then predicted heightened activity of WNT and BMP in the prospective respiratory mesenchyme surrounding the lung bud.
In mouse foregut culture, activating WNT expanded the domain where the respiratory mesenchymal markers Tbx4 and Foxf1 are expressed, while a WNT inhibitor erased Tbx4 expression. Guided by this finding, the authors added WNT stimulation on the fourth day of differentiation of human stem cells and maintained it, which raised TBX4 levels and kept intermediate WNT activity to preserve FOXF1.
🔗 Read original →
OmniSyn builds 2.7‑billion‑entry virtual library for human protein drug design
A preprint describing OmniSyn was posted on bioRxiv, September 6. The computational model takes a protein’s amino‑acid sequence and proposes a candidate molecule together with a step‑by‑step synthesis plan that uses purchasable chemical fragments and reaction types. Unlike many structure‑based methods, OmniSyn relies solely on the sequence, selecting from over 220 000 fragments and 115 common reaction types during training.
The model was tested on 35 proteins not seen during training. For these targets OmniSyn identified exact matches to known active molecules in 0.255% of cases, compared with 0.028% for the next‑best method. When the candidates were analogues of known molecules, the retrosynthesis tool AiZynthFinder found a viable route for 71.92% of them.
For 21 306 human protein sequences OmniSyn generated 2.7 billion target‑molecule entries, each containing a synthetic route and scoring metrics for candidate selection. In a test on 200 proteins, the interaction score distinguished the true target from random proteins (7.64 vs 4.47) and the docking score showed better binding (−6.88 vs −6.20 kcal mol⁻¹).
The library stores, for every protein sequence, a virtual candidate, its synthesis plan
🔗 Read original →
A preprint describing OmniSyn was posted on bioRxiv, September 6. The computational model takes a protein’s amino‑acid sequence and proposes a candidate molecule together with a step‑by‑step synthesis plan that uses purchasable chemical fragments and reaction types. Unlike many structure‑based methods, OmniSyn relies solely on the sequence, selecting from over 220 000 fragments and 115 common reaction types during training.
The model was tested on 35 proteins not seen during training. For these targets OmniSyn identified exact matches to known active molecules in 0.255% of cases, compared with 0.028% for the next‑best method. When the candidates were analogues of known molecules, the retrosynthesis tool AiZynthFinder found a viable route for 71.92% of them.
For 21 306 human protein sequences OmniSyn generated 2.7 billion target‑molecule entries, each containing a synthetic route and scoring metrics for candidate selection. In a test on 200 proteins, the interaction score distinguished the true target from random proteins (7.64 vs 4.47) and the docking score showed better binding (−6.88 vs −6.20 kcal mol⁻¹).
The library stores, for every protein sequence, a virtual candidate, its synthesis plan
🔗 Read original →
bioRxiv
OmniSyn unifies target-aware molecular generation and optimization within a synthesis-native LLM framework across the human proteome
Designing target-specific bioactive molecules with actionable synthesis routes for the human proteome holds enormous potential for expanding therapeutic discovery, but remains a challenge. Existing target-aware generative models often depend on protein structures…
Lipid nanoparticle with cerium oxide, CasRx plasmid, and D4F peptide reduces kidney injury
The authors described a lipid nanoparticle on 6 September that contains three parts: cerium oxide to scavenge reactive oxygen species, a plasmid expressing the CasRx protein guided by a short RNA to target STING transcripts, and the peptide D4F that directs the particle to kidney macrophages. In kidney organoids and a mouse model of transient blood‑flow occlusion, this construct lowered macrophage inflammation, tubular damage, and markers of kidney dysfunction.
Cerium oxide bound excess ROS, the plasmid drove CasRx production that cleaved STING RNA, and D4F facilitated particle accumulation in kidneys and binding to macrophages. In activated mouse macrophages the full construct decreased the proportion of cells showing a fluorescent ROS signal from 71.9% to 24.5%; a second guide RNA gave a similar reduction, while mutating the STING target site partially restored signal levels, confirming the effect depended on CasRx‑STING targeting.
Medium taken from macrophages treated with the nanoparticle reduced death, ROS signaling, and inflammatory gene activity in nearby tubular cells, even though the particles themselves did not reach the tubules, indicating that altered macrophage state can be transmitted via secreted factors. In kidney organoids, D4F increased the fraction of particles‑labeled cells from 9.65% to 27.5%.
In the mouse ischemia‑reperfusion model, two intravenous doses were given two hours before vessel occlusion and 30 minutes after blood flow returned; kidneys were evaluated after one day. Animals receiving the full construct showed the lowest tubular injury, creatinine, and blood urea nitrogen among all groups, with gene‑protein analysis revealing suppressed inflammatory pathways and enhanced oxidative‑stress defense and tissue‑repair programs.
🔗 Read original →
The authors described a lipid nanoparticle on 6 September that contains three parts: cerium oxide to scavenge reactive oxygen species, a plasmid expressing the CasRx protein guided by a short RNA to target STING transcripts, and the peptide D4F that directs the particle to kidney macrophages. In kidney organoids and a mouse model of transient blood‑flow occlusion, this construct lowered macrophage inflammation, tubular damage, and markers of kidney dysfunction.
Cerium oxide bound excess ROS, the plasmid drove CasRx production that cleaved STING RNA, and D4F facilitated particle accumulation in kidneys and binding to macrophages. In activated mouse macrophages the full construct decreased the proportion of cells showing a fluorescent ROS signal from 71.9% to 24.5%; a second guide RNA gave a similar reduction, while mutating the STING target site partially restored signal levels, confirming the effect depended on CasRx‑STING targeting.
Medium taken from macrophages treated with the nanoparticle reduced death, ROS signaling, and inflammatory gene activity in nearby tubular cells, even though the particles themselves did not reach the tubules, indicating that altered macrophage state can be transmitted via secreted factors. In kidney organoids, D4F increased the fraction of particles‑labeled cells from 9.65% to 27.5%.
In the mouse ischemia‑reperfusion model, two intravenous doses were given two hours before vessel occlusion and 30 minutes after blood flow returned; kidneys were evaluated after one day. Animals receiving the full construct showed the lowest tubular injury, creatinine, and blood urea nitrogen among all groups, with gene‑protein analysis revealing suppressed inflammatory pathways and enhanced oxidative‑stress defense and tissue‑repair programs.
🔗 Read original →
PubMed Central (PMC)
D4F‐Functionalized Ceria Nanozyme‐CasRx Platform Suppresses STING and Reprograms the Renal Immune Niche in Acute Kidney Injury
Acute kidney injury (AKI) is sustained by reciprocal amplification of oxidative stress, innate immune signaling, and maladaptive immune–parenchymal crosstalk. Here, we developed a lipid nanoparticle containing a ceria nanozyme and a STING‐targeting ...
Stiff hydrogel boosts early bone formation via PIEZO1 mechanosensor in mice
In mouse experiments, a stiff water‑saturated hydrogel enhanced early bone formation through the mechanosensor PIEZO1. On 6 September, Advanced Science, September 6 published an article showing how a material placed in a bone defect alters cell behavior.
In normal mice with a 1‑mm femoral defect and hind‑limb unloading, a 15% GelMA hydrogel gave more new trabecular bone and collagen by day 7 than a 5% GelMA. This early time point involved five mice per group in the key comparisons.
Cells in the damaged bone contact the matrix, and the authors tested whether matrix stiffness steers cells toward bone or fat programs. Mouse mesenchymal stem cells cultured on gels with 2 and 25 kPa stiffness showed stronger osteogenic markers on the stiffer gel and more lipid droplets on the softer gel under adipogenic induction.
PIEZO1 is a membrane channel that admits calcium when deformed. Blocking mechanosensitive channels with GsMTx4 weakened the stiff‑gel effect, and deleting Piezo1 in mesenchymal progenitors abolished the bone‑program boost on stiff gel.
The authors traced the path from calcium to the gene Stc2, showing that CaMKII inhibition lowered active SP1 and STC2 levels; SP1 binds the Stc2 promoter to drive its transcription. Mutating the promoter reduced SP1 binding and reporter activity, supporting this link.
Even when PIEZO1 was pharmacologically inhibited, adding secreted STC2 shifted cells toward osteogenesis. In a separate experiment, a soft 5% GelMA supplemented with STC2 enhanced early bone‑formation signs in the defect.
In mice lacking Piezo1 in mesenchymal progenitors, neither soft nor stiff gels increased bone formation, and the difference between 5% and 15% GelMA disappeared; these controls were performed without hind‑limb unloading.
🔗 Read original →
In mouse experiments, a stiff water‑saturated hydrogel enhanced early bone formation through the mechanosensor PIEZO1. On 6 September, Advanced Science, September 6 published an article showing how a material placed in a bone defect alters cell behavior.
In normal mice with a 1‑mm femoral defect and hind‑limb unloading, a 15% GelMA hydrogel gave more new trabecular bone and collagen by day 7 than a 5% GelMA. This early time point involved five mice per group in the key comparisons.
Cells in the damaged bone contact the matrix, and the authors tested whether matrix stiffness steers cells toward bone or fat programs. Mouse mesenchymal stem cells cultured on gels with 2 and 25 kPa stiffness showed stronger osteogenic markers on the stiffer gel and more lipid droplets on the softer gel under adipogenic induction.
PIEZO1 is a membrane channel that admits calcium when deformed. Blocking mechanosensitive channels with GsMTx4 weakened the stiff‑gel effect, and deleting Piezo1 in mesenchymal progenitors abolished the bone‑program boost on stiff gel.
The authors traced the path from calcium to the gene Stc2, showing that CaMKII inhibition lowered active SP1 and STC2 levels; SP1 binds the Stc2 promoter to drive its transcription. Mutating the promoter reduced SP1 binding and reporter activity, supporting this link.
Even when PIEZO1 was pharmacologically inhibited, adding secreted STC2 shifted cells toward osteogenesis. In a separate experiment, a soft 5% GelMA supplemented with STC2 enhanced early bone‑formation signs in the defect.
In mice lacking Piezo1 in mesenchymal progenitors, neither soft nor stiff gels increased bone formation, and the difference between 5% and 15% GelMA disappeared; these controls were performed without hind‑limb unloading.
🔗 Read original →
PubMed Central (PMC)
Matrix Stiffness Orchestrates Mesenchymal Stem Cell Lineage Commitment Toward Osteogenesis and Adipogenesis Through the PIEZO1/SP1/STC2…
Mesenchymal stem cells (MSCs) are critical for bone regeneration, and their osteogenic and adipogenic lineage balance is intricately regulated by cellular mechanotransduction. Through single‐cell RNA sequencing reanalysis of the public dataset ...
Hakken model selects three biomedical hypotheses for lab testing
Hakken constructs a map of biomedical entities from publications, linking each pair with a relation type and the time it appeared in the literature. The model compares this structure to article fragments where the entities are mentioned separately and scores hypotheses that are absent from the authors’ working knowledge base. For each candidate it provides a short chain of known facts that explains the high score.
Experts focused on aging, taking 1,386 aging‑related genes; among 959,805 gene pairs Hakken identified 1,543,297 hypotheses above a set threshold. Known interactions were removed, and an automatic selection of the top three candidates per entity left 2,804 hypotheses for further consideration.
Biologists chose three hypotheses with model scores above 0.8 and sent them to an independent contract laboratory. In these experiments a “link” means functional regulation of gene expression: after a cellular intervention the level of messenger RNA (mRNA) or the corresponding protein changes.
For the TP53–BAMBI pair, the drug Nutlin‑3 raised the p53 protein encoded by TP53, and BAMBI mRNA increased by 16% after 24 hours and by 38% after 48 hours.
For the RAF1–TNF pair, the RAF1 inhibitor GW5074 produced a small, repeatable rise in TNF mRNA at 24 hours, while the amount of TNF secreted into the medium stayed below the detection limit.
The third hypothesis, SOAT1–STAT3, received no support: stimulating SOAT1 did not alter STAT3 mRNA or protein levels.
🔗 Read original →
Hakken constructs a map of biomedical entities from publications, linking each pair with a relation type and the time it appeared in the literature. The model compares this structure to article fragments where the entities are mentioned separately and scores hypotheses that are absent from the authors’ working knowledge base. For each candidate it provides a short chain of known facts that explains the high score.
Experts focused on aging, taking 1,386 aging‑related genes; among 959,805 gene pairs Hakken identified 1,543,297 hypotheses above a set threshold. Known interactions were removed, and an automatic selection of the top three candidates per entity left 2,804 hypotheses for further consideration.
Biologists chose three hypotheses with model scores above 0.8 and sent them to an independent contract laboratory. In these experiments a “link” means functional regulation of gene expression: after a cellular intervention the level of messenger RNA (mRNA) or the corresponding protein changes.
For the TP53–BAMBI pair, the drug Nutlin‑3 raised the p53 protein encoded by TP53, and BAMBI mRNA increased by 16% after 24 hours and by 38% after 48 hours.
For the RAF1–TNF pair, the RAF1 inhibitor GW5074 produced a small, repeatable rise in TNF mRNA at 24 hours, while the amount of TNF secreted into the medium stayed below the detection limit.
The third hypothesis, SOAT1–STAT3, received no support: stimulating SOAT1 did not alter STAT3 mRNA or protein levels.
🔗 Read original →
SpringerLink
Link prediction for hypothesis generation: an active curriculum learning infused temporal graph-based approach
Artificial Intelligence Review - Over the last few years Literature-based Discovery (LBD) has regained popularity as a means to enhance the scientific research process. The resurgent interest has...
Jim Greenwood urges longevity movement to partner with biotech groups to shape FDA policy
On September 9, Jim Greenwood told the Alliance for Longevity Initiatives (A4LI) that the longevity movement should work with biotech associations such as BIO and the Alliance for Regenerative Medicine to influence FDA rules. He proposed forming a political action committee (PAC) to repeatedly engage elected officials on a single issue and bringing family members of people with dementia and other age‑related diseases to congressional meetings.
Greenwood served 12 years in the U.S. House of Representatives. He was president and CEO of the Biotechnology Innovation Organization (BIO) from 2005–2020, a period during which BIO’s membership grew to more than >1,000 companies.
He identified the next reauthorization of the Prescription Drug User Fee Act (PDUFA) as a practical window for action, saying A4LI should formulate its requirements in advance and channel them through the associations that already sit at the FDA table. “A4LI will not be sitting at that table… it should work closely with those who are, because they can convey its demands to the FDA,” Greenwood said.
Greenwood tied this political strategy to the likely path of aging‑biology interventions, noting they will first demonstrate efficacy in individual age‑related diseases that already have an approval and reimbursement pathway. In his view, A4LI can simultaneously discuss future drug rules with Congress and the
🔗 Read original →
On September 9, Jim Greenwood told the Alliance for Longevity Initiatives (A4LI) that the longevity movement should work with biotech associations such as BIO and the Alliance for Regenerative Medicine to influence FDA rules. He proposed forming a political action committee (PAC) to repeatedly engage elected officials on a single issue and bringing family members of people with dementia and other age‑related diseases to congressional meetings.
Greenwood served 12 years in the U.S. House of Representatives. He was president and CEO of the Biotechnology Innovation Organization (BIO) from 2005–2020, a period during which BIO’s membership grew to more than >1,000 companies.
He identified the next reauthorization of the Prescription Drug User Fee Act (PDUFA) as a practical window for action, saying A4LI should formulate its requirements in advance and channel them through the associations that already sit at the FDA table. “A4LI will not be sitting at that table… it should work closely with those who are, because they can convey its demands to the FDA,” Greenwood said.
Greenwood tied this political strategy to the likely path of aging‑biology interventions, noting they will first demonstrate efficacy in individual age‑related diseases that already have an approval and reimbursement pathway. In his view, A4LI can simultaneously discuss future drug rules with Congress and the
🔗 Read original →
U.S. Food and Drug Administration
PDUFA VIII: Fiscal Years 2028 – 2032
Information related to FDA’s preparation for the seventh reauthorization of PDUFA.
Nature reports chromosomal damage and mosaic editing in human embryos after base editor PCSK9 tweak
On Nature, September 9, 2026
On September 9, 2026, researchers published a study describing experiments on early human embryos targeting the PCSK9 and HBG genes with the adenine base editor ABE8e-V106W introduced at fertilization. The editor modified all PCSK9 alleles and permitted some embryos to reach the blastocyst stage, yielding embryonic stem cell lines with edited PCSK9.
The paper also notes a rare chromosome break at the target site, chromosomal abnormalities, and mosaic off‑target edits—different cells within the same embryo displayed distinct changes. While standard CRISPR/Cas9 creates double‑strand breaks that in early human embryos are associated with large deletions and loss of chromosomal segments, the adenine base editor produces a single‑strand nick and chemically alters a single DNA base.
PCSK9 and HBG were selected because they have been extensively studied in somatic cell editing, allowing the team to monitor how the embryo repairs such damage. After validating guide RNAs in cellular models, the editor was applied to human zygotes; delivering the editor protein together with its guide RNA enabled some embryos to develop to the blastocyst stage and produce PCSK9‑edited stem cell lines.
mRNA that encodes the editor frequently arrested development at early divisions, a effect the authors attribute to the deaminase activity of the editor independent of the guide RNA. A June 2026 review of the preprint had found no large deletions or chromosomal anomalies after editing, but the September article describes a different damage profile: a rare chromosome break at the target site, chromosomal anomalies, and mosaic off‑target edits near the targets and elsewhere in the genome.
When assessing safety, researchers examine the whole embryonic genome, chromosome status, and whether edits are uniform across cells. The authors conclude that the genomic and developmental consequences currently preclude clinical application of this approach in reproduction.
🔗 Read original →
On Nature, September 9, 2026
On September 9, 2026, researchers published a study describing experiments on early human embryos targeting the PCSK9 and HBG genes with the adenine base editor ABE8e-V106W introduced at fertilization. The editor modified all PCSK9 alleles and permitted some embryos to reach the blastocyst stage, yielding embryonic stem cell lines with edited PCSK9.
The paper also notes a rare chromosome break at the target site, chromosomal abnormalities, and mosaic off‑target edits—different cells within the same embryo displayed distinct changes. While standard CRISPR/Cas9 creates double‑strand breaks that in early human embryos are associated with large deletions and loss of chromosomal segments, the adenine base editor produces a single‑strand nick and chemically alters a single DNA base.
PCSK9 and HBG were selected because they have been extensively studied in somatic cell editing, allowing the team to monitor how the embryo repairs such damage. After validating guide RNAs in cellular models, the editor was applied to human zygotes; delivering the editor protein together with its guide RNA enabled some embryos to develop to the blastocyst stage and produce PCSK9‑edited stem cell lines.
mRNA that encodes the editor frequently arrested development at early divisions, a effect the authors attribute to the deaminase activity of the editor independent of the guide RNA. A June 2026 review of the preprint had found no large deletions or chromosomal anomalies after editing, but the September article describes a different damage profile: a rare chromosome break at the target site, chromosomal anomalies, and mosaic off‑target edits near the targets and elsewhere in the genome.
When assessing safety, researchers examine the whole embryonic genome, chromosome status, and whether edits are uniform across cells. The authors conclude that the genomic and developmental consequences currently preclude clinical application of this approach in reproduction.
🔗 Read original →
Nature
Highly efficient base editing at PCSK9 and normal human embryo development
Nature - Highly efficient base editing at PCSK9 and normal human embryo development
Anthropic's Hubinger Says AI Extinction Risk Over 10% in Next Decade
On September 9, Evan Hubinger, head of AI alignment at Anthropic, replied to a thread by former colleague Jacob Cockson. He agreed with Cockson’s warning about a race toward self‑improving superintelligence and said Anthropic currently lacks a plan to align such a system with human interests.
The same day Jacob Cockson announced his departure from Anthropic. In his thread he described the situation as a race to self‑improving superintelligence, urged researchers to pursue alternative development conditions, and suggested a temporary halt on scaling model capabilities might be needed to avert a global race.
In his reply Hubinger added his personal assessment, stating that he believes the probability of AI causing human extinction in the next ten years is >10%. He said Anthropic is doing
🔗 Read original →
On September 9, Evan Hubinger, head of AI alignment at Anthropic, replied to a thread by former colleague Jacob Cockson. He agreed with Cockson’s warning about a race toward self‑improving superintelligence and said Anthropic currently lacks a plan to align such a system with human interests.
The same day Jacob Cockson announced his departure from Anthropic. In his thread he described the situation as a race to self‑improving superintelligence, urged researchers to pursue alternative development conditions, and suggested a temporary halt on scaling model capabilities might be needed to avert a global race.
In his reply Hubinger added his personal assessment, stating that he believes the probability of AI causing human extinction in the next ten years is >10%. He said Anthropic is doing
🔗 Read original →
X (formerly Twitter)
Evan Hubinger (@EvanHub) on X
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet ha…
AI Agent Calibrates Six‑Qubit Chip, Improves Four of Forty Measurements
GPT‑5.6 Sol performed calibration of four qubits on a new MIT chip on September 8. Researchers intervened in only 4 of 40 target measurements to improve results. The agent, guided by OpenAI’s description of graduate student Beatrice Yankelevich’s work, used the Codex system to control measurements on a separate chip with six superconducting qubits.
Calibration sets the chip before experiments: the outcome of one measurement becomes the
🔗 Read original →
GPT‑5.6 Sol performed calibration of four qubits on a new MIT chip on September 8. Researchers intervened in only 4 of 40 target measurements to improve results. The agent, guided by OpenAI’s description of graduate student Beatrice Yankelevich’s work, used the Codex system to control measurements on a separate chip with six superconducting qubits.
Calibration sets the chip before experiments: the outcome of one measurement becomes the
🔗 Read original →
Philanthropists urged to fund low‑cost gerotherapy trials
On September 7, Fight Aging! published an essay by Reason explaining why testing a cheap therapy can cost far more than the therapy itself. He proposes that wealthy philanthropists fund early clinical programs to generate data. Gerotherapy targets biological aging processes and aims to reduce the risk of multiple age‑related diseases. The compound may be available in clinics, but its effect on aging must be established in a dedicated clinical trial.
In such a trial the participant group and the outcome to be measured—disease, physical function, or a combination—are defined in advance. For interventions that merely slow aging, Reason notes that observation must last at least five years to detect changes in age‑related disease incidence and mortality. He explains the difficulty with patent economics: a patent gives the developer exclusive sales rights and helps recoup R&D costs, whereas a cheap, easily copied therapy has weak patent protection; the trial can still be expensive, and investors have little incentive to pay because any supplier could benefit from the results.
“Interventions that could potentially be used against aging may remain on the market for decades without a sufficiently rigorous attempt to determine whether they actually treat aging and how well.” The essay also considers lowering medical‑regulation costs, medical tourism, and self‑experiment communities. According to the author, these routes broaden access to procedures, but reliable open data still require an organized clinical program. An article by an international geroscience working group explains why such a program needs time and money. Researchers first decide who to include in the trial and which treatment outcome to measure. Slow age‑related decline in physical function is only visible after long observation of large groups.
Scientists have not yet agreed on a universal biomarker—a measurable indicator in lab tests—that could signal aging. Therefore the price of the drug and the price of proof are separate quantities. Seeking early markers that could shorten the validation period is aided by the AFAR FAST program, which analyses samples from completed trials. Reason proposes that wealthy philanthropists fund the initial clinical programs for affordable candidates. In his model, philanthropists finance the trial data, which any supplier of the therapy could use.
🔗 Read original →
On September 7, Fight Aging! published an essay by Reason explaining why testing a cheap therapy can cost far more than the therapy itself. He proposes that wealthy philanthropists fund early clinical programs to generate data. Gerotherapy targets biological aging processes and aims to reduce the risk of multiple age‑related diseases. The compound may be available in clinics, but its effect on aging must be established in a dedicated clinical trial.
In such a trial the participant group and the outcome to be measured—disease, physical function, or a combination—are defined in advance. For interventions that merely slow aging, Reason notes that observation must last at least five years to detect changes in age‑related disease incidence and mortality. He explains the difficulty with patent economics: a patent gives the developer exclusive sales rights and helps recoup R&D costs, whereas a cheap, easily copied therapy has weak patent protection; the trial can still be expensive, and investors have little incentive to pay because any supplier could benefit from the results.
“Interventions that could potentially be used against aging may remain on the market for decades without a sufficiently rigorous attempt to determine whether they actually treat aging and how well.” The essay also considers lowering medical‑regulation costs, medical tourism, and self‑experiment communities. According to the author, these routes broaden access to procedures, but reliable open data still require an organized clinical program. An article by an international geroscience working group explains why such a program needs time and money. Researchers first decide who to include in the trial and which treatment outcome to measure. Slow age‑related decline in physical function is only visible after long observation of large groups.
Scientists have not yet agreed on a universal biomarker—a measurable indicator in lab tests—that could signal aging. Therefore the price of the drug and the price of proof are separate quantities. Seeking early markers that could shorten the validation period is aided by the AFAR FAST program, which analyses samples from completed trials. Reason proposes that wealthy philanthropists fund the initial clinical programs for affordable candidates. In his model, philanthropists finance the trial data, which any supplier of the therapy could use.
🔗 Read original →
PubMed Central (PMC)
Challenges in developing Geroscience trials
Geroscience is becoming a major hope for preventing age-related diseases and loss of function by targeting biological mechanisms of aging. This unprecedented paradigm shift requires optimizing the design of future clinical studies related to aging ...
GPN-Star ranks DNA variants using cross‑species evolutionary constraints
GPN-Star compares a human DNA segment with the same segment across other species to estimate the evolutionary constraint of each single‑letter substitution. It builds a multiple‑sequence alignment of genomes and a phylogenetic tree that shows how closely related the species are. The model was described in Nature, September 9, 2026.
During training GPN-Star masks one DNA letter and predicts it from the surrounding human bases and the corresponding positions in other species, weighting closer relatives more heavily according to branch distances. The output is a probability for each possible base, from which an evolutionary‑constraint score is derived. Separate versions were trained on vertebrate, mammalian and primate alignments.
The vertebrate‑aligned version performed best at identifying substitutions in protein‑coding regions and rare high‑impact variants, whereas the mammalian and primate versions were stronger on regulatory regions that control gene expression. In practice the authors applied GPN‑Star to 34 quantitative traits from the UK Biobank, analysing 161,822 participants of European ancestry.
They added three GPN‑Star scores to the DeepRVAT statistical method, which searches for trait‑associated genes using rare variants. Across three runs the
🔗 Read original →
GPN-Star compares a human DNA segment with the same segment across other species to estimate the evolutionary constraint of each single‑letter substitution. It builds a multiple‑sequence alignment of genomes and a phylogenetic tree that shows how closely related the species are. The model was described in Nature, September 9, 2026.
During training GPN-Star masks one DNA letter and predicts it from the surrounding human bases and the corresponding positions in other species, weighting closer relatives more heavily according to branch distances. The output is a probability for each possible base, from which an evolutionary‑constraint score is derived. Separate versions were trained on vertebrate, mammalian and primate alignments.
The vertebrate‑aligned version performed best at identifying substitutions in protein‑coding regions and rare high‑impact variants, whereas the mammalian and primate versions were stronger on regulatory regions that control gene expression. In practice the authors applied GPN‑Star to 34 quantitative traits from the UK Biobank, analysing 161,822 participants of European ancestry.
They added three GPN‑Star scores to the DeepRVAT statistical method, which searches for trait‑associated genes using rare variants. Across three runs the
🔗 Read original →
Nature
Predicting genome-wide functional constraints with GPN-Star
Nature - GPN-Star, a genomic language model with a phylogeny-aware architecture for whole-genome alignment data, is shown to be a scalable and flexible tool for genetic variant effect prediction...
Blood cell reprogramming erases DNA methylation age marks
Researchers reprogrammed blood cells from 99 healthy donors into induced pluripotent stem cells (iPSCs) to test whether DNA methylation age signatures persist after reprogramming. They published their findings on 8 September.
Starting with peripheral blood mononuclear cells (PBMCs), they obtained 88 iPSC lines from donors aged 22–92 years after quality control. They retained 67 matched pairs of original blood and iPSC samples for direct methylation comparison.
In the original blood, epigenetic clocks showed higher methylation age with older donors, but in the iPSCs the age correlation disappeared across all five clock calculations. A genome‑wide analysis of all measured DNA sites, including those outside the clock sets, confirmed that age‑related signals were present in blood cells but absent in iPSCs.
The team also examined methQTL—statistical links between DNA variants and methylation levels—finding some shared between blood and iPSCs and some specific to each cell state, indicating that genetic influences on methylation differ after reprogramming. Consequently, to study blood‑specific methylation age patterns, the original blood sample is required because fully reprogrammed iPSCs carry a different methylation landscape.
🔗 Read original →
Researchers reprogrammed blood cells from 99 healthy donors into induced pluripotent stem cells (iPSCs) to test whether DNA methylation age signatures persist after reprogramming. They published their findings on 8 September.
Starting with peripheral blood mononuclear cells (PBMCs), they obtained 88 iPSC lines from donors aged 22–92 years after quality control. They retained 67 matched pairs of original blood and iPSC samples for direct methylation comparison.
In the original blood, epigenetic clocks showed higher methylation age with older donors, but in the iPSCs the age correlation disappeared across all five clock calculations. A genome‑wide analysis of all measured DNA sites, including those outside the clock sets, confirmed that age‑related signals were present in blood cells but absent in iPSCs.
The team also examined methQTL—statistical links between DNA variants and methylation levels—finding some shared between blood and iPSCs and some specific to each cell state, indicating that genetic influences on methylation differ after reprogramming. Consequently, to study blood‑specific methylation age patterns, the original blood sample is required because fully reprogrammed iPSCs carry a different methylation landscape.
🔗 Read original →
PromptBio releases preprint on PromptGenie system for biomedical computation tracking
On September 6, the PromptBio team posted a preprint on bioRxiv describing PromptGenie, a system for biomedical data research. The authors explain how it turns a research question into a plan, code, execution, and records that allow verification of the computational workflow. A biomedical question is transformed into output through a series of decisions: what data to use, which method to test a hypothesis, how to perform the calculation, and how to interpret results.
When these decisions reside in separate files, services, and email threads, it is hard to see from a final table or figure which data and choices led to it. In the preprint, the authors call PromptGenie a system that receives a research goal and available context, builds a plan of dependent steps, and adjusts subsequent steps based on intermediate results or errors. Before a major decision, the researcher can review and edit the plan.
The system then invokes a ready-made method or writes code, runs the calculation, and passes the result to the next step. Authors describe a journal where the plan and its modifications, data sources and references, parameters, execution environment details,
🔗 Read original →
On September 6, the PromptBio team posted a preprint on bioRxiv describing PromptGenie, a system for biomedical data research. The authors explain how it turns a research question into a plan, code, execution, and records that allow verification of the computational workflow. A biomedical question is transformed into output through a series of decisions: what data to use, which method to test a hypothesis, how to perform the calculation, and how to interpret results.
When these decisions reside in separate files, services, and email threads, it is hard to see from a final table or figure which data and choices led to it. In the preprint, the authors call PromptGenie a system that receives a research goal and available context, builds a plan of dependent steps, and adjusts subsequent steps based on intermediate results or errors. Before a major decision, the researcher can review and edit the plan.
The system then invokes a ready-made method or writes code, runs the calculation, and passes the result to the next step. Authors describe a journal where the plan and its modifications, data sources and references, parameters, execution environment details,
🔗 Read original →