A deep learning model for predicting next-generation sequencing depth from DNA sequence

Targeted high-throughput DNA sequencing is crucial for genomics and molecular diagnostics. Oligonucleotide probes used for enrichment exhibit varying hybridization kinetics, leading to non-uniform coverage. This paper introduces a deep learning model (DLM) to predict Next-Generation Sequencing (NGS) depth from DNA probe sequences. The DLM, incorporating a bidirectional recurrent neural network, uses DNA nucleotide identities and unpaired nucleotide probabilities. Evaluated on three NGS panels (SNP, lncRNA, and synthetic), the DLM accurately predicts sequencing depth, showing high accuracy in cross-validation and independent testing. The model also effectively predicts DNA hybridization and strand displacement kinetic rate constants.

Jinny X. Zhang, Boyan Yordanov, Alexander Gaunt, Michael X. Wang, Peng Dai, Yuan-Jyue Chen, Kerou Zhang, John Z. Fang, Neil Dalchau, Jiaming Li, Andrew Phillips, David Yu Zhang