svmATAC

svmATAC employs support vector machine (SVM) models to enhance and impute peak signals in Single-cell Assay Transposase Accessible Chromatin sequencing (scATAC-seq) datasets to improve chromatin accessibility profiles for accurate cell type identification.

Key Features:

• Peak enhancement and imputation: Enhances peak signal strength and imputes missing peaks in scATAC-seq data.
• SVM-based modeling: Implements a support vector machine (SVM)-based methodology for signal enhancement and classification.
• Co-accessibility leveraging: Leverages patterns of co-accessibility to inform imputation and signal enhancement.
• Marker-independent identification: Enables cell type identification without reliance on literature-based canonical markers.
• Cross-platform robustness: Demonstrates robustness across datasets from different libraries and sequencing platforms.
• Sparse-data focus: Specifically addresses sparsity caused by lower copy numbers and inherent missing signals in scATAC-seq peaks.

Scientific Applications:

• Cell type classification: Improves accuracy of cell type classification from scATAC-seq datasets.
• Chromatin accessibility profiling: Enhances genome-wide chromatin accessibility profiles derived from scATAC-seq.
• Immune and hematopoietic studies: Applied to analysis of human immune cells, hematopoietic system cells, and peripheral blood mononuclear cells (PBMCs).

Methodology:

Uses a support vector machine (SVM)-based approach to enhance peak signal strength and impute missing scATAC-seq data by leveraging patterns of co-accessibility.