Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN
Abstract Deep learning classification models based on Convolutional Neural Networks (CNNs) are increasingly used in population genetic inference for detecting signatures of natural selection. Prevailing detection methods treat the design of the classifier as a discrete phase, assuming that high clas...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Communications Biology |
| Online Access: | https://doi.org/10.1038/s42003-025-07480-7 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832594489032572928 |
|---|---|
| author | Sjoerd van den Belt Nikolaos Alachiotis |
| author_facet | Sjoerd van den Belt Nikolaos Alachiotis |
| author_sort | Sjoerd van den Belt |
| collection | DOAJ |
| description | Abstract Deep learning classification models based on Convolutional Neural Networks (CNNs) are increasingly used in population genetic inference for detecting signatures of natural selection. Prevailing detection methods treat the design of the classifier as a discrete phase, assuming that high classification accuracy is the sole prerequisite for precise detection. This frequently steers method development toward classification-driven optimizations that can inadvertently impede detection. We present FASTER-NN, a CNN classifier designed specifically for the precise detection of natural selection. It has higher sensitivity than state-of-the-art CNN classifiers while only processing allele frequencies and genomic positions through dilated convolutions to maximize data reuse. As a result, execution time is invariant to the sample size and the chromosome length, creating a highly suitable solution for large-scale, whole-genome scans. Furthermore, FASTER-NN can accurately identify selective sweeps in recombination hotspots, which is a highly challenging detection problem with very limited theoretical treatment to date. |
| format | Article |
| id | doaj-art-7d510ff6529046e8a0fdbed66308a6d2 |
| institution | Kabale University |
| issn | 2399-3642 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Communications Biology |
| spelling | doaj-art-7d510ff6529046e8a0fdbed66308a6d22025-01-19T12:35:19ZengNature PortfolioCommunications Biology2399-36422025-01-018111210.1038/s42003-025-07480-7Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NNSjoerd van den Belt0Nikolaos Alachiotis1University of TwenteUniversity of TwenteAbstract Deep learning classification models based on Convolutional Neural Networks (CNNs) are increasingly used in population genetic inference for detecting signatures of natural selection. Prevailing detection methods treat the design of the classifier as a discrete phase, assuming that high classification accuracy is the sole prerequisite for precise detection. This frequently steers method development toward classification-driven optimizations that can inadvertently impede detection. We present FASTER-NN, a CNN classifier designed specifically for the precise detection of natural selection. It has higher sensitivity than state-of-the-art CNN classifiers while only processing allele frequencies and genomic positions through dilated convolutions to maximize data reuse. As a result, execution time is invariant to the sample size and the chromosome length, creating a highly suitable solution for large-scale, whole-genome scans. Furthermore, FASTER-NN can accurately identify selective sweeps in recombination hotspots, which is a highly challenging detection problem with very limited theoretical treatment to date.https://doi.org/10.1038/s42003-025-07480-7 |
| spellingShingle | Sjoerd van den Belt Nikolaos Alachiotis Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN Communications Biology |
| title | Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN |
| title_full | Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN |
| title_fullStr | Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN |
| title_full_unstemmed | Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN |
| title_short | Fast and accurate deep learning scans for signatures of natural selection in genomes using FASTER-NN |
| title_sort | fast and accurate deep learning scans for signatures of natural selection in genomes using faster nn |
| url | https://doi.org/10.1038/s42003-025-07480-7 |
| work_keys_str_mv | AT sjoerdvandenbelt fastandaccuratedeeplearningscansforsignaturesofnaturalselectioningenomesusingfasternn AT nikolaosalachiotis fastandaccuratedeeplearningscansforsignaturesofnaturalselectioningenomesusingfasternn |