DUET: sequence-only prediction of RNA-small molecule binding sites
- 1Jingdezhen Ceramic University
Abstract
Sequence-only predictors can score RNAs without experimental structures, but continuously updated training pools can reabsorb benchmark chains. Test overlap and within-training redundancy are often handled together, although they affect evaluation differently. DUET combines a relational graph tower with two convolutional sequence towers built on frozen RNA language models. We removed 98 of 388 training and validation records reaching 80% identity to canonical test chains, retained 110 sequence-identical records, and isolated 115 homology components across folds. Across 10 paired seeds, DUET reached 0.240 +/- 0.013 MCC versus 0.192 +/- 0.031 for CoBRA-M3 on the same data (Delta = +0.047; Holm p = 0.0073), with parallel gains in AUC and AUPRC. DUET and GRASSP were statistically indistinguishable on external sets (0.246 versus 0.243 MCC). Real three-dimensional topology improved MCC by +0.037 over shuffled edges and by +0.042 over predicted secondary structure. Increasing the training pool from 80 to 290 records raised DUET by +0.135 MCC. Benchmark evaluation therefore requires separate controls for test overlap and within-training redundancy.
Related
- Canonical EDS1/PAD4 small-molecule binding sites are required for LRR-RP-mediated pattern-triggered immunity
- Functional analysis of small-molecule and protein binding domains using monomeric streptavidin-based two-component biosensors
- LigRNA: High-accuracy prediction of RNA-ligand binding sites via dynamic chemical complementarity and language model encoding
- MtbAIM: AI/ML-guided endpoint-specific prediction of small-molecule activity across drug-susceptible and drug-resistant Mycobacterium tuberculosis
- Nanoarchitecture of the presynaptic cell adhesion molecule PTPσ in relation to prospective vesicle fusion sites and postsynaptic NMDA receptors
BibTeX
@misc{xiong2026duet,
title = {{DUET: sequence-only prediction of RNA-small molecule binding sites}},
author = {K. Xiong and J. Jia},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.28.762699},
url = {https://www.biorxiv.org/content/10.64898/2026.10.28.762699v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - DUET: sequence-only prediction of RNA-small molecule binding sites
AU - K. Xiong
AU - J. Jia
PY - 2026
DA - 2026/10/28
PB - bioRxiv
DO - 10.64898/2026.10.28.762699
UR - https://www.biorxiv.org/content/10.64898/2026.10.28.762699v1
AB - Sequence-only predictors can score RNAs without experimental structures, but continuously updated training pools can reabsorb benchmark chains. Test overlap and within-training redundancy are often handled together, although they affect evaluation differently. DUET combines a relational graph tower with two convolutional sequence towers built on frozen RNA language models. We removed 98 of 388 training and validation records reaching 80% identity to canonical test chains, retained 110 sequence-identical records, and isolated 115 homology components across folds. Across 10 paired seeds, DUET reached 0.240 +/- 0.013 MCC versus 0.192 +/- 0.031 for CoBRA-M3 on the same data (Delta = +0.047; Holm p = 0.0073), with parallel gains in AUC and AUPRC. DUET and GRASSP were statistically indistinguishable on external sets (0.246 versus 0.243 MCC). Real three-dimensional topology improved MCC by +0.037 over shuffled edges and by +0.042 over predicted secondary structure. Increasing the training pool from 80 to 290 records raised DUET by +0.135 MCC. Benchmark evaluation therefore requires separate controls for test overlap and within-training redundancy.
ER -