ALS Cross-Platform Classification - CyTOF and scRNA-seq Data

This dataset accompanies the cross-platform immune-coordination classifier for ALS, from the study "Peripheral immune patterns enable robust cross-platform prediction of ALS onset and progression." It contains mass cytometry (CyTOF) and single-cell RNA sequencing (scRNA-seq) data from peripheral blood mononuclear cells (PBMCs) of ALS patients and healthy controls, used to develop cross-platform machine learning models that classify ALS status and progression rate from immune cell-type coordination. Analysis codeManuscript Dataset 1: als_cytof_annotated.h5adTechnology: Mass cytometry (CyTOF)Samples: 49 samples across 4 batchesGroups: Healthy controls, ALS patients (slow, standard, and fast progression)Cells: 9,396,825Markers: 30 protein markersCell types: 36 immune cell-type labels (obs['cell_type'])Contains: Raw marker intensities (X, float32), cell-type annotations (obs['cell_type']), sample IDs (obs['Sample']), and clinical metadata (progression rate ALSFRS-R/time, progression group [Classifier], sex, age, riluzole, batch).Note: The model's input is per-patient cell-type frequencies, computed from obs['cell_type'] at runtime. Regenerable analysis intermediates (PCA/UMAP embeddings, neighbor graphs) are omitted. Dataset 2: itou_scrna_annotated_v5.h5adTechnology: 10x Genomics scRNA-seqSource: Itou et al. 2024 (GSE244263), re-annotated for cross-platform harmonizationSamples: 40 PBMC samples (30 ALS, 10 controls)Cells: 402,850 (full annotated set; 6,330 cells flagged in obs['excluded'])Genes: 3,000 highly variable genesCell types: 26 harmonized "v5" cell-type labels (obs['cell_type']); intermediate labels preserved as cell_type_v4 and cell_type_baseContains: Log-normalized expression (X, float32), cell-type annotations, QC flag (obs['excluded']), disease state (obs['Disease_state']), sample IDs (obs['Sample_name']), and PCA/Harmony/UMAP embeddings.Use case: Independent scRNA-seq validation cohort for the cross-platform classifier. These datasets enable cross-platform validation of immune signatures in ALS, demonstrating that disease-associated immune coordination patterns identified in CyTOF data generalize to scRNA-seq data. They support reproducible machine learning analyses for ALS patient stratification based on immune p Manuscript: Peripheral immune paatform prediction of ALS onset and progression