ORBIT-FMIB: Tracking Order-Resolved Epistatic Information Through ESM-2
Abstract
Protein foundation models support mutation-effect and structural prediction, but predictive performance alone does not reveal which forms of biological interaction information remain accessible through model depth. We ask whether ESM-2 retains higher-order epistatic information as strongly as first- and second-order information across its representation hierarchy, introducing ORBIT-FMIB, a diagnostic framework combining Walsh-based interaction decomposition with subset-conditioned neural dependence estimation. The method is validated on synthetic landscapes with known interaction structure before being applied to the dense four-site GB1 fitness landscape using frozen ESM-2 representations. An initial production run suggested ESM-2 retains higher-order epistatic information less well than lower-order information (ΔHO−LO=−0.107). An independent replication of the complete measurement grid, under matched GPU hardware and identical critic seeds, substantially reduced this contrast (ΔHO−LO=−0.017), and its sign was unstable across otherwise-defensible evaluation-pairing choices applied to the same trained critics (−0.011 to +0.015). We therefore do not currently have robust evidence that ESM-2 selectively loses higher-order epistatic information, nor that retention is equal across orders; the directional question remains open. The measurement protocol itself, including its documented removal of a positional-subset shortcut in pooled critics, remains validated and is unaffected by this finding. ORBIT-FMIB is offered as a diagnostic framework for probing interaction structure in protein foundation models; this study's own replication result illustrates why such probing requires adequately-powered reproducibility checks before its output is treated as a biological finding.
Related
- From pose to behavior: SABER integrates identity-resolved multi-animal pose tracking with language-model-based behavioral factor discovery
- Tracking demographic changes in individual coral colonies through repeated photogrammetric surveys
- Subcellularly Resolved Single-Cell Embedding Learning with Transcriptomic data, Protein Structure and Localization Information
- Integrating Epistatic Interactions into Genomic Prediction of Growth and Fillet Fat Content in Common Carp
- Scalable phylodynamic inference of mutational and epistatic fitness effects with multi-feature birth-death models
BibTeX
@misc{rahimimovassagh2026orbitfmib,
title = {{ORBIT-FMIB: Tracking Order-Resolved Epistatic Information Through ESM-2}},
author = {Maryam Rahimimovassagh and Ivan Garibay and Niloofar Yousefi},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.00672},
eprint = {2610.00672},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.00672},
}
RIS
TY - UNPB
TI - ORBIT-FMIB: Tracking Order-Resolved Epistatic Information Through ESM-2
AU - Maryam Rahimimovassagh
AU - Ivan Garibay
AU - Niloofar Yousefi
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.00672
UR - https://arxiv.org/abs/2610.00672
AB - Protein foundation models support mutation-effect and structural prediction, but predictive performance alone does not reveal which forms of biological interaction information remain accessible through model depth. We ask whether ESM-2 retains higher-order epistatic information as strongly as first- and second-order information across its representation hierarchy, introducing ORBIT-FMIB, a diagnostic framework combining Walsh-based interaction decomposition with subset-conditioned neural dependence estimation. The method is validated on synthetic landscapes with known interaction structure before being applied to the dense four-site GB1 fitness landscape using frozen ESM-2 representations. An initial production run suggested ESM-2 retains higher-order epistatic information less well than lower-order information (Δ_HO−LO=−0.107). An independent replication of the complete measurement grid, under matched GPU hardware and identical critic seeds, substantially reduced this contrast (Δ_HO−LO=−0.017), and its sign was unstable across otherwise-defensible evaluation-pairing choices applied to the same trained critics (−0.011 to +0.015). We therefore do not currently have robust evidence that ESM-2 selectively loses higher-order epistatic information, nor that retention is equal across orders; the directional question remains open. The measurement protocol itself, including its documented removal of a positional-subset shortcut in pooled critics, remains validated and is unaffected by this finding. ORBIT-FMIB is offered as a diagnostic framework for probing interaction structure in protein foundation models; this study's own replication result illustrates why such probing requires adequately-powered reproducibility checks before its output is treated as a biological finding.
ER -