JSOM

JSOM aligns and integrates related single-cell datasets by constructing two low-dimensional discretized maps that jointly evolve to match similar clusters while preserving each dataset's topological structure for downstream clustering and meta-analysis.

Key Features:

• Map construction: Constructs two low-dimensional discretized representations ("maps") that evolve jointly according to both input datasets.
• Joint evolution alignment: Joint evolution aligns similar clusters across datasets while preserving topological relationships within each dataset.
• Heterogeneity handling: Accommodates differing numbers of measured parameters and distinct data distributions across datasets.
• Batch-effect mitigation: Addresses batch effects commonly observed within datasets produced using identical methodologies.
• SOM-based dimensionality reduction: Implements an advanced variation of the Self-organizing Map (SOM) for dimensionality reduction and visualization.
• Downstream analysis support: Preserves topology to support clustering and meta-analysis of aligned datasets.
• Single-cell technology compatibility: Applied to flow cytometry, mass cytometry, and single-cell RNA sequencing datasets.

Scientific Applications:

• Flow cytometry: Aligns related clusters across flow cytometry datasets while preserving intrinsic topology to enable integrated analysis.
• Mass cytometry: Aligns and integrates mass cytometry datasets across experiments to facilitate comparative and meta-analyses.
• Single-cell RNA sequencing: Aligns single-cell RNA sequencing datasets generated by different technologies or batches to support combined clustering and interpretation.

Methodology:

Constructs two low-dimensional discretized maps that jointly evolve according to both input datasets using an advanced variation of the Self-organizing Map (SOM), aligning similar clusters while preserving topological structure.