DeepSV

DeepSV is a software tool that utilizes deep learning to identify structural variations, specifically long deletions, in DNA sequence data. It extends Google's approach to identifying single nucleotide polymorphisms (SNPs) to more complex genetic variations.

Key features of DeepSV:

1. Novel visualization method: DeepSV introduces a new way of visualizing sequence reads that capture multiple sources of information relevant to long deletions.

2. Handling noisy training data: The tool implements techniques to work with noisy training data, which is common in genomic datasets.

3. Deep learning model: DeepSV trains a deep learning model using the visualized sequence reads to call deletions based on the learned patterns.

Topic

DNA polymorphism;DNA structural variation;Machine learning;Imaging;Sequencing

Detail

  • Operation: SNP detection;Variant calling;Visualisation;Split read mapping

  • Software interface: Command-line interface

  • Language: Python

  • License: Not stated

  • Cost: Free of charge

  • Version name: -

  • Credit: Beijing Natural Science Foundation, Fundamental Research Funds for the Central Universities, Research projects on biomedical transformation of China-Japan Friendship Hospital, and US National Science Foundation.

  • Input: -

  • Output: -

  • Contact: Jingyang Gao gaojy@mail.buct.edu.cn

  • Collection: -

  • Maturity: -

Publications

  • DeepSV: accurate calling of genomic deletions from high-throughput sequencing data using deep convolutional neural network.
  • Cai L, et al. DeepSV: accurate calling of genomic deletions from high-throughput sequencing data using deep convolutional neural network. DeepSV: accurate calling of genomic deletions from high-throughput sequencing data using deep convolutional neural network. 2019; 20:665. doi: 10.1186/s12859-019-3299-y
  • https://doi.org/10.1186/S12859-019-3299-Y
  • PMID: 31830921
  • PMC: PMC6909530

Download and documentation

< Back to DB search