PAUSE: principled feature attribution for unsupervised gene expression analysis
Abstract As interest in using unsupervised deep learning models to analyze gene expression data has grown, an increasing number of methods have been developed to make these models more interpretable. These methods can be separated into two groups: post hoc analyses of black box models through featur...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2023-04-01
|
| Series: | Genome Biology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13059-023-02901-4 |
| _version_ | 1827961301299625984 |
|---|---|
| author | Joseph D. Janizek Anna Spiro Safiye Celik Ben W. Blue John C. Russell Ting-I Lee Matt Kaeberlin Su-In Lee |
| author_facet | Joseph D. Janizek Anna Spiro Safiye Celik Ben W. Blue John C. Russell Ting-I Lee Matt Kaeberlin Su-In Lee |
| author_sort | Joseph D. Janizek |
| collection | DOAJ |
| description | Abstract As interest in using unsupervised deep learning models to analyze gene expression data has grown, an increasing number of methods have been developed to make these models more interpretable. These methods can be separated into two groups: post hoc analyses of black box models through feature attribution methods and approaches to build inherently interpretable models through biologically-constrained architectures. We argue that these approaches are not mutually exclusive, but can in fact be usefully combined. We propose PAUSE ( https://github.com/suinleelab/PAUSE ), an unsupervised pathway attribution method that identifies major sources of transcriptomic variation when combined with biologically-constrained neural network models. |
| first_indexed | 2024-04-09T16:24:32Z |
| format | Article |
| id | doaj.art-e28d031d40c24e3483d38d0e92b91ca7 |
| institution | Directory Open Access Journal |
| issn | 1474-760X |
| language | English |
| last_indexed | 2024-04-09T16:24:32Z |
| publishDate | 2023-04-01 |
| publisher | BMC |
| record_format | Article |
| series | Genome Biology |
| spelling | doaj.art-e28d031d40c24e3483d38d0e92b91ca72023-04-23T11:18:56ZengBMCGenome Biology1474-760X2023-04-0124113010.1186/s13059-023-02901-4PAUSE: principled feature attribution for unsupervised gene expression analysisJoseph D. Janizek0Anna Spiro1Safiye Celik2Ben W. Blue3John C. Russell4Ting-I Lee5Matt Kaeberlin6Su-In Lee7Paul G. Allen School of Computer Science and Engineering, University of WashingtonPaul G. Allen School of Computer Science and Engineering, University of WashingtonRecursion PharmaceuticalsDepartment of Pathology, University of WashingtonDepartment of Pathology, University of WashingtonDepartment of Pathology, University of WashingtonDepartment of Pathology, University of WashingtonPaul G. Allen School of Computer Science and Engineering, University of WashingtonAbstract As interest in using unsupervised deep learning models to analyze gene expression data has grown, an increasing number of methods have been developed to make these models more interpretable. These methods can be separated into two groups: post hoc analyses of black box models through feature attribution methods and approaches to build inherently interpretable models through biologically-constrained architectures. We argue that these approaches are not mutually exclusive, but can in fact be usefully combined. We propose PAUSE ( https://github.com/suinleelab/PAUSE ), an unsupervised pathway attribution method that identifies major sources of transcriptomic variation when combined with biologically-constrained neural network models.https://doi.org/10.1186/s13059-023-02901-4TranscriptomicsGene expressionDeep learningExplainable AIUnsupervised learningFeature attribution |
| spellingShingle | Joseph D. Janizek Anna Spiro Safiye Celik Ben W. Blue John C. Russell Ting-I Lee Matt Kaeberlin Su-In Lee PAUSE: principled feature attribution for unsupervised gene expression analysis Genome Biology Transcriptomics Gene expression Deep learning Explainable AI Unsupervised learning Feature attribution |
| title | PAUSE: principled feature attribution for unsupervised gene expression analysis |
| title_full | PAUSE: principled feature attribution for unsupervised gene expression analysis |
| title_fullStr | PAUSE: principled feature attribution for unsupervised gene expression analysis |
| title_full_unstemmed | PAUSE: principled feature attribution for unsupervised gene expression analysis |
| title_short | PAUSE: principled feature attribution for unsupervised gene expression analysis |
| title_sort | pause principled feature attribution for unsupervised gene expression analysis |
| topic | Transcriptomics Gene expression Deep learning Explainable AI Unsupervised learning Feature attribution |
| url | https://doi.org/10.1186/s13059-023-02901-4 |
| work_keys_str_mv | AT josephdjanizek pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT annaspiro pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT safiyecelik pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT benwblue pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT johncrussell pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT tingilee pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT mattkaeberlin pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis AT suinlee pauseprincipledfeatureattributionforunsupervisedgeneexpressionanalysis |