Gene function prediction from bulk coexpression is bounded by cell-type-level signal
- 1BC Cancer Research Center
Abstract
It is widely accepted in genomics that coexpression of RNA transcripts suggests a commonality of function. This intuition is explicitly leveraged in machine learning methods that predict gene function, where it is often combined with other features such as protein interactions and sequence similarity. For example, including coexpression data from human tissue expression boosts performance for predicting Gene Ontology annotations. However, the biological underpinnings of this observation have not been well-investigated. Building on earlier results from our group, in this work we show that gene function is predictable from coexpression substantially because it reflects differences in expression between cell types, and these differences are also intrinsic to the ground truth labels. Using simulations and analyses of real data, we show that variance in the cellular composition of bulk samples impacts function learnability and attribute this to cell type marker gene content in the GO terms. We further show that cell type profiles, where the relationship between gene expression and cell type is made transparent, are effective for predicting gene function while increasing interpretability. These results indicate that function prediction models trained on bulk coexpression are largely limited to cell-type-level resolution rather than fine-grained biochemical function, with direct consequences for how such predictions should be interpreted.
Related
- Inflammatory Subphenotypes for Fluid Management in Acute Respiratory Distress Syndrome: A Subgroup-Level Signal, not an Individual Prescription
- A De Novo, Intragenic Modifier in Cis Rescues Scn8a-N1768D Epilepsy and Defines a Therapeutic Window Bounded by Gain- and Loss-of-Function
- OMICON: a community resource for studying gene coexpression networks in normal and neoplastic human brain samples
- How Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR Model
- Odorant receptor coexpression and multi-expression in the dengue mosquito
BibTeX
@misc{adrianhamazaki2026gene,
title = {{Gene function prediction from bulk coexpression is bounded by cell-type-level signal}},
author = {A. Adrian-Hamazaki and P. Pavlidis},
howpublished = {bioRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.31.748380},
url = {https://www.biorxiv.org/content/10.64898/2026.08.31.748380v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Gene function prediction from bulk coexpression is bounded by cell-type-level signal
AU - A. Adrian-Hamazaki
AU - P. Pavlidis
PY - 2026
DA - 2026/08/31
PB - bioRxiv
DO - 10.64898/2026.08.31.748380
UR - https://www.biorxiv.org/content/10.64898/2026.08.31.748380v1
AB - It is widely accepted in genomics that coexpression of RNA transcripts suggests a commonality of function. This intuition is explicitly leveraged in machine learning methods that predict gene function, where it is often combined with other features such as protein interactions and sequence similarity. For example, including coexpression data from human tissue expression boosts performance for predicting Gene Ontology annotations. However, the biological underpinnings of this observation have not been well-investigated. Building on earlier results from our group, in this work we show that gene function is predictable from coexpression substantially because it reflects differences in expression between cell types, and these differences are also intrinsic to the ground truth labels. Using simulations and analyses of real data, we show that variance in the cellular composition of bulk samples impacts function learnability and attribute this to cell type marker gene content in the GO terms. We further show that cell type profiles, where the relationship between gene expression and cell type is made transparent, are effective for predicting gene function while increasing interpretability. These results indicate that function prediction models trained on bulk coexpression are largely limited to cell-type-level resolution rather than fine-grained biochemical function, with direct consequences for how such predictions should be interpreted.
ER -