Mamba-GRN: A Mamba-inspired framework for no-overlap held-out regulatory edge prediction
Computational Biology and Chemistry · 2027-02 · 126 · 109369
DOI: 10.1016/j.compbiolchem.2026.109369 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computational Biology and Chemistry, 2027-02. Verfasst von HaiLong Wu und 2 weiteren Autor:innen. Beteiligte Einrichtung: East China University of Technology (ECUT), Nanchang 330013, Jiangxi, China. Electronic address: Glyne2905@outlook.com. Hauptthemen laut MEDLINE: Gene Regulatory Networks, Algorithms, Computational Biology. Die Arbeit nennt 12 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Major Discipline Academic and Technical Leaders Training Program of Jiangxi Province, National Science and Technology Major Project.
Abstract (Originalsprache)
Reliable evaluation of gene regulatory network (GRN) inference requires strict separation between training and held-out regulatory edges, particularly for negative edges in sparse networks. We present Mamba-GRN, a compact Mamba-inspired gene representation and edge-decoding framework evaluated under a corrected no-overlap protocol in which validation and test negatives are excluded from the training-negative pool. The implemented encoder combines expression-derived features and learnable gene-identity embeddings with residual blocks composed of layer normalization, linear expansion, depthwise one-dimensional convolution, GELU activation, and linear projection; it does not implement a selective-scan state-space recurrence. Across seven non-tiny GSD datasets and three random seeds, the full model achieved mean AUROC 0.6225, AUPRC 0.4754, and Precision@P 0.4444, compared with 0.5708/0.4022/0.4444 for GENIE3 and 0.5671/0.4446/0.3810 for GRNBoost2. Paired mean improvements over the mature tree-based baselines were positive, but Holm-adjusted Wilcoxon tests did not reach the 0.05 threshold; the revised analysis therefore reports effect estimates, bootstrap confidence intervals, and win rates without claiming universal statistical superiority. Sensitivity analyses showed broadly stable performance across 1:1, 2:1, and 5:1 training-negative ratios, while larger representation dimensions improved mean performance at increased parameter cost. In an independent K562 Perturb-seq benchmark with 2284 aligned genes and 20,795 perturbation-response associations, source-matched hard-negative evaluation yielded AUROC 0.7582 ± 0.0071 and AUPRC 0.6114 ± 0.0124. The pretrained frozen backbone provided only a modest, seed-dependent advantage over a randomly initialized frozen backbone. These results support Mamba-GRN as a controlled framework for held-out edge recovery, while limiting the claims to the evaluated networks, candidate-edge setting, and functional perturbation-response associations.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- East China University of Technology (ECUT), Nanchang 330013, Jiangxi, China. Electronic address: Glyne2905@outlook.com.
- East China University of Technology (ECUT), Nanchang 330013, Jiangxi, China.
- East China University of Technology (ECUT), Nanchang 330013, Jiangxi, China; Nanchang Key Laboratory of Brain-Computer Interface and Intelligent Rehabilitation Equipment, Nanchang, 330013, Jiangxi, China.
Förderung
- National Natural Science Foundation of China
- Major Discipline Academic and Technical Leaders Training Program of Jiangxi Province
- National Science and Technology Major Project
- National Major Science and Technology Projects of China
Zitieren
APA
Wu, H., Wu, Z., & Liu, X. (2027). Mamba-GRN: A Mamba-inspired framework for no-overlap held-out regulatory edge prediction. Computational Biology and Chemistry, 126, 109369. https://doi.org/10.1016/j.compbiolchem.2026.109369
Vancouver
Wu H, Wu Z, Liu X. Mamba-GRN: A Mamba-inspired framework for no-overlap held-out regulatory edge prediction. Computational Biology and Chemistry. 2027;126:109369. doi:10.1016/j.compbiolchem.2026.109369
BibTeX
@article{Wu2027,
author = {Wu, HaiLong and Wu, ZhiMo and Liu, XiaoQiong},
title = {Mamba-GRN: A Mamba-inspired framework for no-overlap held-out regulatory edge prediction},
journal = {Computational Biology and Chemistry},
year = {2027},
volume = {126},
pages = {109369},
doi = {10.1016/j.compbiolchem.2026.109369}
}
EUVIMED