Life sciences · Preprint

NEAT-POCKET: Pocket-Conditioned Autoregressive 3D Molecular Generation with a Neighborhood-Guided Set Transformer

arXiv · September 4, 2026

Preliminary Why this matters to you Expert discussion

Early or partial results. Treat as a signal, not a conclusion.

AI Summary

NEAT-POCKET is a newly described computational autoregressive model for pocket-conditioned 3D molecular generation that shows competitive benchmarking performance and faster sampling than baselines on CrossDocked and SPINDR datasets. The work is methodological and remains unvalidated experimentally; it represents an algorithmic advance in silico but has no direct evidence of utility in drug discovery or biological efficacy.

What was studied

Computational benchmark study. Intervention: NEAT-POCKET model for pocket-conditioned 3D molecular generation, trained and evaluated on CrossDocked and SPINDR datasets. Compared with: Existing baseline methods for structure-based molecular generation.

Key findings

NEAT-POCKET achieves competitive structure-based generation performance on CrossDocked and SPINDR datasets Model samples substantially faster than existing baselines Enables pocket-conditioned fragment completion relevant to lead optimization and scaffold elaboration

Safety

Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.

Who this may help

The source did not state who this applies to in practice.

Plain-language summary

An early-stage computational model for AI-driven molecular generation with benchmarking on datasets, but no validation in wet lab, clinical, or prospective drug discovery settings.

Study design

As stated by the source record.

MethodsComputational benchmark study

InterventionNEAT-POCKET model for pocket-conditioned 3D molecular generation, trained and evaluated on CrossDocked and SPINDR datasets

ComparatorExisting baseline methods for structure-based molecular generation

Caveats

No quantitative performance metrics (e.g., success rate, RMSD, binding affinity predictions) are reported in the abstract.
No experimental validation: generated molecules have not been synthesized or tested for binding or biological activity.
Not peer reviewed; this is a preprint on arXiv.
Generalizability to novel protein targets or clinical drug candidates is not assessed.

Evidence assessment

Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.

Study design

limited

Computational benchmark on existing datasets; no experimental or prospective validation design.

Directness

limited

Surrogate endpoint (computational performance); no evidence of functional binding or biological activity.

Consistency

insufficient

Source does not report consistency across independent datasets or replication beyond two named benchmarks.

Risk of bias

insufficient

No description of dataset selection criteria, validation strategy, or potential confounders in benchmarking.

Sample adequacy

insufficient

Source does not report sample size, number of molecules generated, or molecules evaluated.

What is missing. This record has no reported figures. That is a gap in the analysis, not a judgement about the study.

Source abstract

AI-driven de novo molecular design offers a promising route to accelerate early-stage drug discovery by generating novel ligands directly within target protein binding pockets. We present NEAT-POCKET, a pocket-conditioned extension of the autoregressive NEAT model for 3D molecular generation. NEAT-POCKET generates molecules atom by atom in protein pocket environments while preserving atom permutation invariance and explicitly modeling hydrogen atoms. Benchmarks on the CrossDocked and SPINDR datasets show that NEAT-POCKET achieves competitive structure-based generation performance while sampling substantially faster than existing baselines. Beyond full-molecule generation, NEAT-POCKET naturally enables pocket-conditioned fragment completion, a task directly relevant to lead optimization and scaffold elaboration. These results position NEAT-POCKET as a fast, flexible, and practical framework for structure-based drug design.

Follow from this

Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.

Johannes KirchmairDaniel RoseRoxane Axel JacobThierry Langer