Cheminformatics
10 articles
Machine learning tiers PFAS liver hazard, and liver spheroid gene data back it up
- An XGBoost QSAR model trained on high-throughput bioactivity data sorts PFASs into four tiers of liver-related hazard.
- Public transcriptomes from human liver spheroids exposed to 19 PFASs showed response patterns that varied across those tiers.
- Lipid metabolism disruption and cellular stress dominated the signatures, with FABP1, PLIN2 and GDF15 responding across exposures.
Tiers from a QSAR model inherit the weaknesses of their training labels, and high-throughput bioactivity assays rarely capture slow liver injury. PFAS persistence also means spheroid concentrations may sit far from real accumulation, so checking tiers against measured tissue exposure and serum half-life matters. FABP1, PLIN2 and GDF15 look promising as readouts, but stress markers are notoriously non-specific.
Pairwise-difference QSAR with quantile aggregation extrapolates beyond known potency
- Standard QSAR struggles to predict analogs more potent than anything in its training set, which is exactly the lead-optimization question.
- Pairwise-Difference QSAR learns activity changes from descriptor differences between compound pairs, then aggregates the predictions.
- On ChEMBL-derived benchmarks across about a hundred targets with the most potent compounds held out, a low-quantile aggregation beat mean and kNN aggregation, direct QSAR and Chemprop.
Every optimization team has felt this: the model ranks the series fine, then fails at the top end where you actually need it. Holding out the most potent decile is the right stress test. My caution is that public ChEMBL tails differ from a live program's SAR cliffs, so I'd trial it prospectively on one of our own series before trusting it to prioritize synthesis.
Breaking the Potency Barrier: A Practical Recipe for QSAR Extrapolation ↗
Benchmark asks whether structure-based methods rescue DEL-trained ML out of distribution
- A recent NeurIPS competition showed top ML models trained on DEL data failing to generalize to out-of-distribution chemical space; this study tests whether structural modeling can bridge the gap.
- It benchmarks ML, Glide docking, GALigandDock and Boltz-2 cofolding on three biologically diverse targets screened against DELs built in several synthesis formats.
- ML does well in-distribution; the paper's focus is how each approach holds up on new chemistry.
DEL-trained models that shine on random splits and collapse on new chemotypes is exactly what bites when you try to seed a different series from a DEL campaign. Three targets is a small panel, but a like-for-like comparison with physics-based methods on identical data is useful. I'd use it to calibrate how much to trust ML triage on scaffolds the library never sampled.
In silico building-block selection widens the diversity of a 403,200-member peptide-like DEL
- A three-cycle, peptide-like DEL of 403,200 compounds was built, with a chemical space model of 4,774 commercial amino acids used to find under-represented building blocks.
- Affinity selections were run against three targets to compare library performance.
- Selection against carbonic anhydrase II recovered the known sulfonamide pharmacophore, supporting the computational design approach.
Recovering the sulfonamide on CA II is the expected sanity check, not proof of lead-like coverage. Peptide-like DELs often return big, polar hits, so what I'd judge is how many selected compounds sit in usable property space off-DNA, and how the two other targets behaved.
Modeling-led design yields a pan-serotype dengue NS4B inhibitor candidate
- Comparative models of the structurally unresolved NS4B across all four dengue serotypes reveal a conserved transmembrane pocket that fits known resistance mutations.
- Docking of the JNJ-series inhibitors fed predictive field-based 3D-QSAR models, which were used to screen bioisosteres and in-house compounds.
- The top candidate, BNJ-495, was extremely potent against serotype 2 with a high selectivity index, and its conserved binding contacts suggest broad-spectrum activity.
A homology-model plus 3D-QSAR workflow that delivers a prospectively validated hit against a membrane protein with no structure is rare, and that is what makes this worth a look. Before I trust 'pan-serotype', I'd want to see the other three serotypes tested and whether the JNJ-series liabilities carried over.
Structure-Guided Design of a Subnanomolar Dengue Virus NS4B Inhibitor ↗
Scope library: a chemical space between fragments and full molecules for tractable hits
- Scope defines a chemical space between fragments and elaborated low-molecular-weight library compounds, emphasizing compound quality, ligand efficiency and selected property ranges to keep hits tractable for medicinal chemistry.
- Scope ring systems (ring-X-ring) organize the space and guide intentional functionalization toward low- to mid-micromolar potency; a roughly 10k-compound set suits medium-throughput SPR, native MS, DSF and biochemical screens.
- Case studies on trypanosomal CLK1, a 14-3-3/ERα molecular glue project and SARS-CoV-2 Mpro show comparable primary hit rates to diversity sets, with better ligand efficiency after counter-screening.
Honestly, the pitch I like here is hits you can actually start working on: a library sitting between fragments and full-size compounds, built around ligand efficiency, with ring-X-ring cores that tell you where to functionalize next. A 10k set fits SPR, native MS and DSF, so a smaller group can run it. The open question is whether primary hit rates on par with diversity sets, with better LE after counter-screening, hold up beyond three targets, and how much the scaffold constraint narrows novelty.
Rule-Based SMILES Editing Tool Aids Fragment-to-Lead Optimization
- A new method systematically edits SMILES strings—swapping atom types, resizing or opening rings, and grafting fragments—to generate analog ideas for fragment-to-lead optimization.
- First implemented as a KNIME workflow tested on fragments from multiple programs, it's now also available as open-source Python code paired with AutoDock Vina docking.
- This non-learned, rule-based approach complements AI/ML generative design, offering a practical option when training data for a target is scarce.
Honestly this is what I'd do by hand anyway—ring contraction, bioisostere swaps, fragment grafting—just automated and exhaustive instead of me sketching ten ideas on a whiteboard. For a scarce-data fragment series, I'd trust this over a generative model I can't sanity-check. Free KNIME/Python beats begging IT for another SaaS license too.
Fragment Optimization─Simple SMILES String Manipulation Strategies to Optimize a Hit or Fragment ↗
Docking Showdown: Physics-Based, Co-Folding, and Hybrid Models Face Off
- Compares physics-based docking, deep-learning co-folding, and hybrid rescoring methods on how well they generalize to unfamiliar protein targets and chemical scaffolds.
- Builds on the recent Runs-n-Poses benchmark, which found that classical physics-based docking held an edge over several co-folding tools on truly novel targets.
- Assesses pose accuracy and ranking performance to offer practical guidance on choosing a docking strategy based on target novelty.
Physics-based docking still beating co-folding tools on genuinely novel targets is the finding I keep coming back to — these deep-learning models are pattern-matching PDB coverage, not solving folding physics, and it shows the moment a target has no close analog. Until a co-folding tool proves itself on a zero-analog case, I'm running both and trusting the classical scorer as the tiebreaker.
Multiparadigm Benchmark of Molecular Docking: From Physics to Co-Folding and Hybrid Models ↗
Benchmark Study Reveals Blind Spots in AI-Based Virtual Screening for Keap1-Nrf2
- A new benchmarking study examines a virtual-screening pipeline that chains structure-based co-folding models with learned-affinity scoring, an increasingly common setup for computational hit-finding.
- The cascade was applied to the Keap1-Nrf2 protein-protein interaction, a notoriously hard-to-drug target due to its flat, polar interface.
- Instead of trusting only the final ranked output, the authors assessed each stage separately against experimentally meaningful signal.
- They identify specific points where a stage's score diverges from that signal, a failure mode that can mislead downstream compound prioritization.
- The paper proposes per-stage control checks practitioners can run before relying on a cascade's final hit list, framing the work as a methodology and reliability study rather than a report of a new inhibitor.
Keap1-Nrf2 is exactly the flat, polar PPI where a co-folding model will happily spit out a confident-looking pose and the affinity scorer will rank it #1 for the wrong reasons — so a paper that pries the cascade apart stage-by-stage is more useful to me than another "we found a hit" claim. What I actually want out of this: a checklist to run on my own vendor or in-house pipeline before I let its top 20 dictate what goes on the synthesis queue. If the co-folding confidence and the affinity score are quietly diverging, that's exactly the failure mode that burns weeks on compounds that never had real signal behind them. No new chemical matter here, no binding mode to chase — just a reminder that for a target this treacherous, the ranked list is not the QC step. Filed under "read before ordering," not "read for the inhibitor."
Many Virtual Screening 'Improvements' May Not Be Real, Pre-Registered Study Finds
- The preprint asks a foundational methods question: given the size of a typical structure-based virtual screening benchmark, how large a difference in AUROC or enrichment between two protocols can actually be distinguished from statistical noise.
- Rather than assuming benchmark comparisons are inherently reliable, the authors first directly measure these resolution limits.
- They then pre-registered six candidate interventions to docking/scoring protocols and tested each against the measured resolution limits, rather than reporting post-hoc 'wins.'
- The pre-registration design is meant to guard against selective reporting, a known issue where small, benchmark-specific gains get presented as general improvements.
- This is a methodology/rigor study, not a new-inhibitor or target-specific paper; it addresses how virtual screening improvements should be evaluated.
- Practical implication: many published claims of improved virtual screening performance may fall within the noise floor of typical benchmarks, so hit-finding teams should treat single-benchmark improvement claims cautiously.
Every VS vendor pitch should have to clear this paper's bar. If a 'protocol improvement' can't beat the measured noise floor of AUROC/enrichment on a benchmark our size, it's not an improvement — it's a coin flip dressed up in a webinar slide. Six pre-registered interventions is a high bar; realistically most of us won't pre-register anything before a campaign, but I can at least ask collaborators to report effect sizes against a resolution limit, not just a headline AUROC bump. Sharpest implication for me: the co-folding hype (Boltz-2 and friends) needs the same treatment — if those models are just pattern-matching PDB-adjacent chemotypes, their 'wins' over classic docking/FEP on public benchmarks could be memorization, not generalization. I'm not swapping FEP for a black box on this evidence alone.