SMRT

SMRT implements a progressive computational approach for SNP calling and haplotype assembly from Single Molecular Sequencing (SMS) data to produce phased variant calls despite SMS's higher error rates.

Key Features:

• Data type: Processes Single Molecular Sequencing (SMS) reads that can cover approximately 90% of chromosomal positions.
• Progressive approach: Implements a progressive computational method tailored to mitigate SMS's higher error rate for SNP identification and haplotype construction.
• Scalability: Capable of processing large datasets such as >200 million non-N bases on Chromosome 1 with millions of reads across >100 blocks.
• Block processing: Handles blocks that can contain upwards of 2 million bases and average around 3,000 SNP sites per block.
• SNP density per sample: Reported average SNP sites per block are 3,378 for NA12878 and 5,736 for NA24385.
• Performance metrics: Reported false discovery rates are ~15.7% (NA12878) and ~16.5% (NA24385), false negative rate ~11.0%, and switch errors 7.26 (NA12878) and 5.21 (NA24385).

Scientific Applications:

• SNP calling: Produces variant calls from high-error-rate SMS reads for single-nucleotide polymorphism analysis.
• Haplotype assembly: Generates phased haplotypes from SMS data for downstream phasing analyses.
• Genetic diversity and ancestry studies: Provides phased variant and haplotype information suitable for analyses of genetic diversity and ancestry.

Methodology:

Implements a progressive computational approach specifically tailored for SMS data and is reported to process large-scale inputs (e.g., >200 million non-N bases on Chromosome 1 across millions of reads and >100 blocks).