Back

GeroQubit: a lightweight, honesty-first de-novo design platform for geroscience-native small molecules with calibrated uncertainty

k, D.; Swetha, H.

2026-06-11 bioinformatics
10.64898/2026.06.07.730687 bioRxiv
Show abstract

Computational molecule generation has outpaced its own credibility. We present GeroQubit, a GPU-free de-novo design platform that organizes candidates along a target x tissue x hallmark model and reports every signal alongside its measured baseline. We treat our tissue aging-signature readout as a mechanistic structural prior that we explicitly disclose is not validated against lifespan, and we surface efficacy only through a structure-to-lifespan k-NN whose weak but real signal (leave-one-out {rho} {approx} 0.145) is wrapped in empirically-calibrated conformal intervals (90% target, 90.3% measured coverage). On a held-out retrospective recovery of [~]1,940 ChEMBL binders against decoys, the score reaches ROC-AUC 0.945 with [~]20x enrichment at 1% (BEDROC 0.91) and survives a scaffold-disjoint split -- yet we report that it collapses to near-random (AUC 0.62) on genuinely novel chemotypes. Molecules are assembled reaction-first, so every candidate carries a verified synthetic route and atom-level synthon provenance; ADMET is handled as a multi-objective Pareto problem. We frame the disclosed weak signals and the hard-case failures not as flaws but as the honest, decision-useful output the fields own critics demand.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

1
Nature
○ 645 papers in training set
Top 1%
11.7%
Similar papers in this journal
2
Nature Communications
○ 5641 papers in training set
Top 22%
7.7%
Similar papers in this journal
3
Cell Systems
○ 201 papers in training set
Top 0.5%
6.6%
Similar papers in this journal
4
Nature Methods
○ 385 papers in training set
Top 2%
6.6%
Similar papers in this journal
5
Science
● 477 papers in training set
Top 1%
6.1%
Similar papers in this journal
6
eLife
● 5828 papers in training set
Top 23%
5.4%
7
PLOS Computational Biology
● 1863 papers in training set
Top 8%
4.2%
8
Patterns
○ 78 papers in training set
Top 0.5%
3.4%
50% of probability mass above
9
Nature Aging
○ 60 papers in training set
Top 0.6%
3.2%
10
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 18%
3.1%
11
Bioinformatics
◐ 1204 papers in training set
Top 6%
2.6%
12
Cell Genomics
○ 172 papers in training set
Top 2%
2.4%
13
Nature Biotechnology
○ 172 papers in training set
Top 2%
2.3%
14
Briefings in Bioinformatics
◐ 354 papers in training set
Top 4%
2.1%
15
Nature Machine Intelligence
○ 70 papers in training set
Top 1%
1.9%
16
Nature Genetics
○ 286 papers in training set
Top 3%
1.9%
17
Scientific Reports
○ 3612 papers in training set
Top 54%
1.7%
18
Molecular Systems Biology
○ 162 papers in training set
Top 2%
1.6%
19
Cell
○ 431 papers in training set
Top 7%
1.5%
20
Nature Chemical Biology
○ 119 papers in training set
Top 2%
1.5%
21
npj Aging
○ 22 papers in training set
Top 0.2%
1.4%
22
Communications Chemistry
○ 48 papers in training set
Top 0.9%
1.1%
23
Science Advances
● 1243 papers in training set
Top 28%
1.0%
24
Nature Biomedical Engineering
○ 47 papers in training set
Top 2%
0.8%
25
PLOS ONE
● 5266 papers in training set
Top 60%
0.8%
26
Nature Metabolism
○ 69 papers in training set
Top 2%
0.8%
27
Neuron
○ 337 papers in training set
Top 5%
0.8%
28
ACS Central Science
● 71 papers in training set
Top 2%
0.6%
29
iScience
○ 1154 papers in training set
Top 42%
0.6%
30
Journal of Chemical Information and Modeling
● 238 papers in training set
Top 3%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.