ProSolo

ProSolo implements a probabilistic model to call single nucleotide variants (SNVs) and quantify genotyping uncertainty from multiple displacement amplification (MDA)-amplified single-cell DNA sequencing data by jointly modeling a bulk sample to account for amplification bias and errors.

Key Features:

• Probabilistic SNV calling: Applies a probabilistic model to call single nucleotide variants (SNVs) from single-cell DNA sequencing data.
• Empirical amplification-bias model: Incorporates an empirical model of MDA amplification bias to adjust allele observations.
• Genotyping uncertainty quantification: Quantifies genotyping uncertainty for single-cell calls.
• Joint single-cell and bulk modeling: Simultaneously analyzes an MDA-amplified single cell sample and a bulk sequencing sample from the same cell population to account for amplification errors.
• Genotype imputation: Performs biologically relevant imputation of missing genotypes in single cells using information from the bulk sample.
• False discovery rate control: Implements a novel, flexible approach to control the false discovery rate (FDR) for variant calls and related inferences.
• Artifact identification: Identifies single-cell-specific artifacts such as allele dropout induced by whole-genome amplification.

Scientific Applications:

• Single-cell SNV detection: Detects SNVs in MDA-amplified single-cell DNA sequencing datasets.
• Single-cell genotyping and imputation: Provides genotypes for single cells and imputes missing genotypes using bulk data.
• Genomic heterogeneity analysis: Supports analysis of mutational profiles across individual cells to study genomic heterogeneity.
• Amplification artifact identification: Enables detection and characterization of amplification-induced errors such as allele dropout.

Methodology:

Uses a probabilistic model that incorporates an empirical MDA amplification-bias model to jointly analyze MDA-amplified single-cell and bulk sequencing reads, quantify genotyping uncertainty, perform genotype imputation, control FDR, and identify allele-dropout artifacts.