In-silico biological discovery with large perturbation models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915218864472064 |
|---|---|
| author | Miladinovic, Djordje Höppe, Tobias Chevalley, Mathieu Georgiou, Andreas Stuart, Lachlan Mehrjou, Arash Bantscheff, Marcus Schölkopf, Bernhard Schwab, Patrick |
| author_facet | Miladinovic, Djordje Höppe, Tobias Chevalley, Mathieu Georgiou, Andreas Stuart, Lachlan Mehrjou, Arash Bantscheff, Marcus Schölkopf, Bernhard Schwab, Patrick |
| contents | Data generated in perturbation experiments link perturbations to the changes they elicit and therefore contain information relevant to numerous biological discovery tasks -- from understanding the relationships between biological entities to developing therapeutics. However, these data encompass diverse perturbations and readouts, and the complex dependence of experimental outcomes on their biological context makes it challenging to integrate insights across experiments. Here, we present the Large Perturbation Model (LPM), a deep-learning model that integrates multiple, heterogeneous perturbation experiments by representing perturbation, readout, and context as disentangled dimensions. LPM outperforms existing methods across multiple biological discovery tasks, including in predicting post-perturbation transcriptomes of unseen experiments, identifying shared molecular mechanisms of action between chemical and genetic perturbations, and facilitating the inference of gene-gene interaction networks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_23535 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | In-silico biological discovery with large perturbation models Miladinovic, Djordje Höppe, Tobias Chevalley, Mathieu Georgiou, Andreas Stuart, Lachlan Mehrjou, Arash Bantscheff, Marcus Schölkopf, Bernhard Schwab, Patrick Machine Learning Quantitative Methods Methodology Data generated in perturbation experiments link perturbations to the changes they elicit and therefore contain information relevant to numerous biological discovery tasks -- from understanding the relationships between biological entities to developing therapeutics. However, these data encompass diverse perturbations and readouts, and the complex dependence of experimental outcomes on their biological context makes it challenging to integrate insights across experiments. Here, we present the Large Perturbation Model (LPM), a deep-learning model that integrates multiple, heterogeneous perturbation experiments by representing perturbation, readout, and context as disentangled dimensions. LPM outperforms existing methods across multiple biological discovery tasks, including in predicting post-perturbation transcriptomes of unseen experiments, identifying shared molecular mechanisms of action between chemical and genetic perturbations, and facilitating the inference of gene-gene interaction networks. |
| title | In-silico biological discovery with large perturbation models |
| topic | Machine Learning Quantitative Methods Methodology |
| url | https://arxiv.org/abs/2503.23535 |