GLOM TGITO: multiscale protein modelling and multiview landscape-guided sequence discovery
- 1AlphaMol Science Ltd.
Abstract
Motivation: Offline protein design must seek high-activity variants while retaining support for their predicted performance. The challenge is to explore multiple promising regions without confining the search to average-performing sequences. We test whether multiscale representations and local evidence can guide this search. Results: GLOM-TGITO links a global-local fluorescence predictor to multiview landscape-guided optimization. Across 31,434 measured ppluGFP2 sequences, GLOM achieved random/depth-holdout RMSEs of 0.170/0.247 versus 0.247/0.324 for the strongest SaProt-650M controls. Averaged across three training seeds, 96.3%/97.3% of prediction-ranked Top 100 test variants were measured bright (log10 fluorescence >=4.2). Search from 900 peak regions yielded 2,594 distinct candidate sequences. The selected Top 100 showed 2.9-8.2% lower mean representation deviation than four comparator pipelines; 38 versus 11-16 met joint novelty-reliability filters. Deviation identified measured test errors >0.1 (AUROC 0.773). The first molecular-dynamics run of the selected Top 30 identified smaller surface hydrophobic patches in 20 candidates relative to their starts, with representative variants showing more persistent chromophore contacts or narrower torsional fluctuations.
Related
- sabinaMBM: An R package for Multiscale Bayesian species distribution Modelling using INLA
- A multiscale pharmacokinetic-Boolean modelling framework for ibrutinib resistance in a Chronic Lymphocytic Leukaemia case study
- Causal discovery and inference in COVID-19: A multiscale longitudinal analysis using mixtures of DAGs
- From single-sequence structure prediction to protein fitness landscape through a composable, epistasis-aware mutation atlas
- Multiscale modelling of drug-host-pathogen interaction: quantifying drug and immune contributions to treatment response
BibTeX
@misc{xiong2026glom,
title = {{GLOM TGITO: multiscale protein modelling and multiview landscape-guided sequence discovery}},
author = {Y. Xiong and Y. K. Tirta and S. Wang and Z. Hu and W. Ma and S. Yuan},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.22.760886},
url = {https://www.biorxiv.org/content/10.64898/2026.10.22.760886v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - GLOM TGITO: multiscale protein modelling and multiview landscape-guided sequence discovery
AU - Y. Xiong
AU - Y. K. Tirta
AU - S. Wang
AU - Z. Hu
AU - W. Ma
AU - S. Yuan
PY - 2026
DA - 2026/10/22
PB - bioRxiv
DO - 10.64898/2026.10.22.760886
UR - https://www.biorxiv.org/content/10.64898/2026.10.22.760886v1
AB - Motivation: Offline protein design must seek high-activity variants while retaining support for their predicted performance. The challenge is to explore multiple promising regions without confining the search to average-performing sequences. We test whether multiscale representations and local evidence can guide this search. Results: GLOM-TGITO links a global-local fluorescence predictor to multiview landscape-guided optimization. Across 31,434 measured ppluGFP2 sequences, GLOM achieved random/depth-holdout RMSEs of 0.170/0.247 versus 0.247/0.324 for the strongest SaProt-650M controls. Averaged across three training seeds, 96.3%/97.3% of prediction-ranked Top 100 test variants were measured bright (log10 fluorescence >=4.2). Search from 900 peak regions yielded 2,594 distinct candidate sequences. The selected Top 100 showed 2.9-8.2% lower mean representation deviation than four comparator pipelines; 38 versus 11-16 met joint novelty-reliability filters. Deviation identified measured test errors >0.1 (AUROC 0.773). The first molecular-dynamics run of the selected Top 30 identified smaller surface hydrophobic patches in 20 candidates relative to their starts, with representative variants showing more persistent chromophore contacts or narrower torsional fluctuations.
ER -