Correcting gradient-based interpretations of deep neural networks for genomics

Post hoc attribution methods can provide insights into the learned patterns from deep neural networks (DNNs) trained on high-throughput functional genomics data. However, in practice, their resultant attribution maps can be challenging to interpret due to spurious importance scores for seemingly arb...

Full description

Saved in:
Bibliographic Details
Published in:Genome Biology 2023-05, Vol.24 (1), p.109-109, Article 109
Main Authors: Majdandzic, Antonio, Rajesh, Chandana, Koo, Peter K
Format: Article
Language:English
Subjects:
Attribution methods
Deep learning
DNA
Euclidean space
Explainable AI
Gene mapping
genome
Genomics
Learning
Model interpretability
Neural networks
Neural Networks, Computer
Nucleotides
Regulatory genomics
Short Report
Transcription factors
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Post hoc attribution methods can provide insights into the learned patterns from deep neural networks (DNNs) trained on high-throughput functional genomics data. However, in practice, their resultant attribution maps can be challenging to interpret due to spurious importance scores for seemingly arbitrary nucleotides. Here, we identify a previously overlooked attribution noise source that arises from how DNNs handle one-hot encoded DNA. We demonstrate this noise is pervasive across various genomic DNNs and introduce a statistical correction that effectively reduces it, leading to more reliable attribution maps. Our approach represents a promising step towards gaining meaningful insights from DNNs in regulatory genomics.
ISSN:1474-760X
1474-7596
1474-760X
DOI:10.1186/s13059-023-02956-3