Genomic
Whole-genome (WGS), whole-exome (WES), single-cell DNA, and targeted clinical panels: somatic/germline variants, structural variants, copy-number alterations, and clonal architecture.
Virtual LabBiobankBuilt and growing
A disease-specific biobank of biomedical data.
Harmonised genomic, transcriptomic, single-cell, spatial, proteomic, and cytometric cohorts—paired with curated clinical-research annotations from global public archives and proprietary partner biobanks. Every dataset is reprocessed from raw reads through the exact same deterministic, containerised pipelines used for client analyses and indexed by disease ontology directly inside the BioMedsAI workspace alongside your own projects.
Public cohortsProprietary cohortsIllustration of how the biobank is organised.
What it holds
Transcriptomic
Bulk RNA-seq and isoform-level expression across diseased, malignant, and matched healthy reference tissues.
Single-cell & Spatial
Donor-resolved scRNA-seq, snRNA-seq, CITE-seq, scATAC-seq, scDNA, and spatial transcriptomics integrated across multi-institutional cohorts.
Proteomic
High-plex surface-marker panels (ADT/AbSeq), mass spectrometry (PRIDE), and secretome measurements co-registered with their corresponding transcriptomes.
Flow & Mass Cytometry
Multi-parameter flow cytometry and CyTOF datasets, gated reproducibly via automated pipelines and linked directly to matched donor sequencing samples.
Clinical-research
Harmonised treatment response, survival, cytogenetic, and pathology annotations governed strictly under source-specific consent and licensing. Zero PHI.
Where the data come from
Global Public Archives (INSDC)
Raw reads and functional genomics cohorts from NCBI GEO & SRA (US), EMBL-EBI ENA & ArrayExpress/BioStudies (Europe), DDBJ/GEA (Japan), and CNCB GSA (China).
Single-Cell & Spatial Atlases
High-resolution reference cohorts from the Human Cell Atlas (HCA), CZ CELLxGENE, Broad Single Cell Portal, HuBMAP, EMBL-EBI SCEA, and GTEx/ENCODE.
Oncology & Disease Consortia
Clinical-genomic and perturbation cohorts from NCI GDC (TCGA, TARGET, BeatAML), MMRF CoMMpass, ICGC-ARGO, cBioPortal, DepMap, and LINCS.
Controlled-Access Biobanks
Authorized genotype-phenotype and clinical cohorts (including NIH dbGaP, EMBL-EBI EGA, JGA, and UK Biobank) ingested strictly under approved Data Use Certifications (DUCs).
Proprietary & Partner Cohorts
Internal sequencing runs and licensed biobank cohorts held in cryptographically segregated, tenant-isolated containers under strict contractual scope with full audit logging.
Organised by disease
Cohorts of one disease sit together, harmonised, so a finding can be set against every independent cohort that bears on it.
Hematologic malignancies
- AML
- ALL
- MDS
- MPN
- CML
- CMML
- CLL
- Multiple myeloma
- Hodgkin lymphoma
- Non-Hodgkin lymphoma
Pre-malignant and benign blood disorders
- Clonal hematopoiesis (CHIP)
- MGUS
- Aplastic anemia
- Sickle cell disease
Solid tumours
- Breast
- Lung
- Colorectal
- Prostate
- Pancreatic
- Ovarian
- Liver
- Kidney
- Glioblastoma
- Melanoma
Immune and inflammatory
- Rheumatoid arthritis
- Lupus
- Inflammatory bowel disease
- Multiple sclerosis
Neurodegenerative
- Alzheimer's disease
- Parkinson's disease
Reference
- Healthy tissue and blood
- Healthy ageing
More diseases are added as new cohorts are reprocessed.
How a cohort enters the biobank.
- 1
Collect & Curate
Ingest global public cohorts by disease indication alongside controlled-access and proprietary datasets strictly under their governing licences and consents.
- 2
Reprocess from Raw Reads
Execute uniform, version-locked pipelines directly from raw FASTQ/BAM files—never stitching together pre-processed count matrices generated by incompatible upstream tools.
- 3
Harmonise & Batch-Correct
Enforce standardised disease/cell ontologies (Mondo, Cell Ontology) and explicit donor, batch, and modality modelling so independent cohorts compare quantitatively.
- 4
Keep the Negatives
Targets and biomarkers that fail independent cohort replication are permanently logged and surfaced—never quietly dropped to inflate discovery claims.
Ask about access.
Tell us the disease, data types and question you have in mind. Access is opening inside the workspace; partners contributing data are welcome too.