KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens
Abstract Characterising context‐dependent gene functions is crucial for understanding the genetic bases of health and disease. To date, inference of gene functions from large‐scale genetic perturbation screens is based on ad hoc analysis pipelines involving unsupervised clustering and functional enr...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Springer Nature
2020-03-01
|
Series: | Molecular Systems Biology |
Subjects: | |
Online Access: | https://doi.org/10.15252/msb.20199083 |
_version_ | 1797285110748282880 |
---|---|
author | Heba Z Sailem Jens Rittscher Lucas Pelkmans |
author_facet | Heba Z Sailem Jens Rittscher Lucas Pelkmans |
author_sort | Heba Z Sailem |
collection | DOAJ |
description | Abstract Characterising context‐dependent gene functions is crucial for understanding the genetic bases of health and disease. To date, inference of gene functions from large‐scale genetic perturbation screens is based on ad hoc analysis pipelines involving unsupervised clustering and functional enrichment. We present Knowledge‐ and Context‐driven Machine Learning (KCML), a framework that systematically predicts multiple context‐specific functions for a given gene based on the similarity of its perturbation phenotype to those with known function. As a proof of concept, we test KCML on three datasets describing phenotypes at the molecular, cellular and population levels and show that it outperforms traditional analysis pipelines. In particular, KCML identified an abnormal multicellular organisation phenotype associated with the depletion of olfactory receptors, and TGFβ and WNT signalling genes in colorectal cancer cells. We validate these predictions in colorectal cancer patients and show that olfactory receptors expression is predictive of worse patient outcomes. These results highlight KCML as a systematic framework for discovering novel scale‐crossing and context‐dependent gene functions. KCML is highly generalisable and applicable to various large‐scale genetic perturbation screens. |
first_indexed | 2024-03-07T17:58:26Z |
format | Article |
id | doaj.art-237a9479d0044ece8ae00efad08f0afd |
institution | Directory Open Access Journal |
issn | 1744-4292 |
language | English |
last_indexed | 2024-03-07T17:58:26Z |
publishDate | 2020-03-01 |
publisher | Springer Nature |
record_format | Article |
series | Molecular Systems Biology |
spelling | doaj.art-237a9479d0044ece8ae00efad08f0afd2024-03-02T11:20:03ZengSpringer NatureMolecular Systems Biology1744-42922020-03-01163n/an/a10.15252/msb.20199083KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screensHeba Z Sailem0Jens Rittscher1Lucas Pelkmans2Department of Engineering Science Institute of Biomedical Engineering University of Oxford Oxford UKDepartment of Engineering Science Institute of Biomedical Engineering University of Oxford Oxford UKDepartment of Molecular Life Sciences University of Zurich Zurich SwitzerlandAbstract Characterising context‐dependent gene functions is crucial for understanding the genetic bases of health and disease. To date, inference of gene functions from large‐scale genetic perturbation screens is based on ad hoc analysis pipelines involving unsupervised clustering and functional enrichment. We present Knowledge‐ and Context‐driven Machine Learning (KCML), a framework that systematically predicts multiple context‐specific functions for a given gene based on the similarity of its perturbation phenotype to those with known function. As a proof of concept, we test KCML on three datasets describing phenotypes at the molecular, cellular and population levels and show that it outperforms traditional analysis pipelines. In particular, KCML identified an abnormal multicellular organisation phenotype associated with the depletion of olfactory receptors, and TGFβ and WNT signalling genes in colorectal cancer cells. We validate these predictions in colorectal cancer patients and show that olfactory receptors expression is predictive of worse patient outcomes. These results highlight KCML as a systematic framework for discovering novel scale‐crossing and context‐dependent gene functions. KCML is highly generalisable and applicable to various large‐scale genetic perturbation screens.https://doi.org/10.15252/msb.20199083cell morphology and microenvironmentCRISPR and siRNA screeningfunctional genomicshigh content screeningolfactory receptors |
spellingShingle | Heba Z Sailem Jens Rittscher Lucas Pelkmans KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens Molecular Systems Biology cell morphology and microenvironment CRISPR and siRNA screening functional genomics high content screening olfactory receptors |
title | KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens |
title_full | KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens |
title_fullStr | KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens |
title_full_unstemmed | KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens |
title_short | KCML: a machine‐learning framework for inference of multi‐scale gene functions from genetic perturbation screens |
title_sort | kcml a machine learning framework for inference of multi scale gene functions from genetic perturbation screens |
topic | cell morphology and microenvironment CRISPR and siRNA screening functional genomics high content screening olfactory receptors |
url | https://doi.org/10.15252/msb.20199083 |
work_keys_str_mv | AT hebazsailem kcmlamachinelearningframeworkforinferenceofmultiscalegenefunctionsfromgeneticperturbationscreens AT jensrittscher kcmlamachinelearningframeworkforinferenceofmultiscalegenefunctionsfromgeneticperturbationscreens AT lucaspelkmans kcmlamachinelearningframeworkforinferenceofmultiscalegenefunctionsfromgeneticperturbationscreens |