Author: itembalabantu

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice. The 30-second version What. On FoldBench, a leakage-controlled independent benchmark, AlphaFold3 (AF3) leads on most tasks — but its score splits sharply by axis. It is robust on protein monomers (LDDT 0.88) and…

The landscape of biology foundation models: do they actually beat the old specialized baselines?

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice. The 30-second version What. Protein, genome and cell “foundation models” (bio-FMs) are usually sold with headlines like “simulating millions of years of evolution.” The firm’s one question is narrower: on a specific task,…

The ATTR-CM commercial map — revenue and narrative are decoupled: stabilizers own the market, siRNA becomes a blockbuster, and the real bottlenecks are the add-on outcome threshold and the price-value gap

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All revenue, prescription, pricing and reduction figures are attributed to the company’s own results (press release / SEC), to the ICER report, or to market-analysis reporting. The acoramidis-versus-tafamidis real-world comparison…

Hard outcomes in ATTR-CM — mortality and CV events are settled within each trial, but “which modality is better” is undeterminable because there is not a single head-to-head RCT

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. Every hard-outcome figure is attributed to its own trial’s primary paper or press release; the eplontersen miss and monotherapy signal are attributed to the AstraZeneca/Ionis press release of 2026-07-09 (press-release…

The TTR silencers — siRNA vutrisiran succeeded, ASO eplontersen missed, but the deciding variable was the trial population, not the modality

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All knockdown, hazard-ratio and subgroup figures are attributed to the sponsoring trial, journal or company; several eplontersen numbers are a company press release rather than peer-reviewed data (noted inline), with…

Stabilizers in ATTR-CM — tafamidis and acoramidis built the first hard-outcome standard of care, but the “next-generation” increment is not proven mortality superiority

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All efficacy figures are attributed to the sponsoring trial or company (ATTR-ACT to Pfizer/NEJM 2018; ATTRibute-CM to BridgeBio/NEJM 2024). Cross-trial comparisons are not head-to-head. The “near-complete (≥90%) stabilization” claim is…

The ATTR cardiomyopathy landscape — four modalities target one protein, but the siRNA-success/ASO-miss split is cross-trial, not head-to-head

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All reduction and outcome figures are attributed to the sponsoring trial or company; some are company press releases rather than peer-reviewed data (noted inline). The eplontersen result below is a…

AI protein design’s validation reality — high in-silico scores, low wet-lab hit rates, and why an honest success-rate benchmark is still missing

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. This is the bottleneck layer of the AI-protein-design series, where its central question resolves. Self-reported in-silico success rates and improvement factors (from original authors and vendors) are separated throughout from…

Function — where AI protein design actually ships binding, catalysis and neutralization, and where it still fails. A ladder: binders on top, enzymes in the middle, antibodies at the bottom

—

by

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All success-rate, affinity and kcat/KM figures are attributed to the primary paper (Nature/Science) or company IR, with the media tier noted inline. Reported success rates are relative to the candidates…

The generative engine for protein design — backbone diffusion + inverse folding + a self-consistency filter that only checks folding, not function

—

by