Learning Genomic Sequence Representations using Graph Neural Networks over De Bruijn Graphs
Abstract
The rapid expansion of genomic sequence data calls for new methods to achieve robust sequence representations. Existing techniques often neglect intricate structural details, emphasizing mainly contextual information. To address this, we developed k-mer embeddings that merge contextual and structural string information by enhancing De Bruijn graphs with structural similarity connections. Subsequently, we crafted a self-supervised method based on Contrastive Learning that employs a heterogeneous Graph Convolutional Network encoder and constructs positive pairs based on node similarities. Our embeddings consistently outperform prior techniques for Edit Distance Approximation and Closest String Retrieval tasks.
Cite
@article{arxiv.2312.03865,
title = {Learning Genomic Sequence Representations using Graph Neural Networks over De Bruijn Graphs},
author = {Kacper Kapuśniak and Manuel Burger and Gunnar Rätsch and Amir Joudaki},
journal= {arXiv preprint arXiv:2312.03865},
year = {2023}
}
Comments
Poster at "NeurIPS 2023 New Frontiers in Graph Learning Workshop (NeurIPS GLFrontiers 2023)"