Related Experiment Video
Updated: Jan 2, 2026
09:40
Generation of Genomic Deletions in Mammalian Cell Lines via CRISPR/Cas9
Published on: January 3, 2015
96.4K
DeepSV: accurate calling of genomic deletions from high-throughput sequencing data using deep convolutional neural
Lei Cai1, Yufeng Wu2, Jingyang Gao3
1Department of Information Science and Technology, Beijing University of Chemical Technology, Beijing, People's Republic of China.
BMC Bioinformatics
|December 14, 2019
Summary
DeepSV uses deep learning to identify long deletions, a complex genetic variation. This novel approach enhances accuracy and efficiency in genetic variation calling compared to existing methods.
Area of Science:
- Genomics
- Bioinformatics
- Machine Learning
Background:
- Accurate genetic variation calling is crucial in genomics.
- Existing methods primarily focus on single nucleotide polymorphisms (SNPs).
- Deep learning shows promise for SNP calling by visualizing sequence data.
Purpose of the Study:
- To investigate the application of deep learning for calling complex genetic variations, specifically structural variations (SVs).
- To develop and evaluate DeepSV, a deep learning-based method for identifying long deletions from sequence reads.
Main Methods:
- Developed DeepSV, a deep learning approach for structural variation detection.
- Introduced a novel visualization method for sequence reads tailored to long deletions.
- Implemented techniques to handle noisy training data.
- Trained a deep neural network model on visualized sequence data.
Main Results:
- DeepSV accurately identifies long deletions from sequence reads.
- The method demonstrates superior accuracy and efficiency compared to existing deletion calling techniques.
- Performance validated on data from the 1000 Genomes Project.
Conclusions:
- Deep learning is effective for calling complex genetic variations beyond SNPs.
- DeepSV represents a significant advancement in structural variation detection.
