NanoCaller

NanoCaller applies deep learning to detect single nucleotide polymorphisms (SNPs) and small insertions and deletions (indels) from long-read sequencing data (Oxford Nanopore and PacBio) to improve variant calling in difficult-to-map genomic regions.

Key Features:

• Deep Learning Framework: Employs a deep neural network to analyze long-read sequencing data from Oxford Nanopore and PacBio for SNP and indel detection.
• Long-Range Haplotype Information: Incorporates long-range haplotype information to enhance SNP detection across larger DNA segments.
• Phasing and Local Realignment: Phases long reads with detected SNPs and performs local realignment to improve indel calling precision.
• Performance Evaluation: Evaluated across eight human genomes and experimentally validated 41 novel variants in a benchmarking genome.

Scientific Applications:

• Variant discovery in difficult-to-map regions: Detects SNPs and small indels in genomic regions that are challenging for short-read sequencing technologies.
• Human genome benchmarking and novel variant discovery: Supports benchmarking and discovery workflows, as demonstrated by validation of novel variants.
• Studies of complex genomic landscapes: Enables analysis of genetic variation across larger, complex genomic segments using long-read data.

Methodology:

SNP detection using deep learning that analyzes long-range haplotype information; phasing of long reads with detected SNPs; indel calling via local realignment.