Pythiomics

Pythiomics is a multi-omics database developed and curated by Pythia Biosciences, designed to bring fragmented public omics data into a single, unified, and standardized resource. Public omics datasets are often scattered across multiple repositories and stored in inconsistent formats, units, and metadata standards — creating significant barriers to accessibility and harmonization. Pythiomics addresses these challenges by combining state-of-the-art AI techniques for metadata harmonization and cell type prediction with meticulous manual curation and quality control, delivering a reliable data resource for biopharmaceutical companies and research institutions.

The platform is built to accelerate data analysis, data integration, and data-driven drug discovery. Pythiomics is interactively accessible through the C-DIAM Multi-Omics Studio platform, providing an easy-to-use graphical interface alongside advanced machine learning algorithms and analysis workflows — making it usable by all scientists who want to leverage public omics data.

Core Data Quality Principles

• Standard Operating Process (SOP): Every dataset follows a defined framework that outlines QC criteria, processing steps, naming conventions, and guidance for handling edge cases.
• Harmonized metadata: Diseases, tissues, treatments, cell types, genders, sampling ages, and additional attributes are all mapped to standardized vocabularies to ensure consistency across studies.
• Rigorous quality control: Automated checks handle scale-level validation, while manual review ensures accuracy and catches issues that algorithms may miss.
• Documentation and versioning: Every processing step is fully recorded — including input data types, QC thresholds, rationale, and code used — with complete version histories to support transparency and traceability.

Omics Data Types and Coverage

• Pythiomics currently incorporates data from more than 10,000 multi-omics datasets drawn from multiple public databases.
• Currently available omics types include Bulk RNA-seq, Single-cell RNA-seq, Proteomics, Visium Spatial Transcriptomics, and CosMx spatial data.
• Additional data types coming soon include Visium HD and Xenium, Metabolomics, ATAC-Seq and CITE-Seq, CHIP-Seq, and Mutation and GWAS data.

Single-Cell RNA-Seq Database Highlights

• Single-cell RNA-seq data is a key focus area within Pythiomics.
• The single-cell database contains over 143 million cells sourced from 7,736 donors.
• It spans 1,745 datasets covering 318 diseases, providing broad coverage for disease-focused research.

Interactive Exploration and Analytics via C-DIAM Multi-Omics Studio

• Pythiomics is hosted within the C-DIAM Multi-Omics Studio platform, enabling interactive exploration through a graphical user interface.
• Users can access a rich package of state-of-the-art machine learning algorithms and analysis workflows directly within the platform.
• API delivery is also available, allowing programmatic access to Pythiomics data for integration into existing research pipelines and workflows.

Pythiomics is available to biopharmaceutical companies and research institutions via a request access process. Scientists can also explore API delivery options to integrate Pythiomics data directly into their computational environments.