MOSAIC Window

MOSAIC Window is the part of the MOSAIC cancer atlas that researchers outside the consortium can apply for. MOSAIC (Multi-Omics Spatial Atlas in Cancer) is a multi-center study led by Owkin that profiles tumor samples with spatial and single-cell methods next to histology and sequencing, to find cancer subtypes, drug targets and biomarkers. The Window subset lets academic groups test their own analyses on the same kind of data. The full cohort is a separate entry (MOSAIC).

Composition

The subset covers 60 patients in five cancer types: 15 with muscle-invasive bladder cancer treated by cystectomy, 15 with ovarian cancer, 10 with glioblastoma, 10 with diffuse large B-cell lymphoma and 10 with mesothelioma. Each cancer type is a separate EGA dataset. They hold spatial transcriptomics, single-cell RNA sequencing, bulk RNA sequencing, whole exome sequencing, H&E histology images and clinical data, plus processed expression tables and variant calls. The EGA study summary gives 50 patients, but its five datasets add up to 60, the number the preprint reports.

Acquisition

The samples are formalin-fixed, paraffin-embedded tumor blocks. The sources do not say which hospitals contributed the 60 patients. According to the preprint, all participating centers scanned the H&E slides on a Leica Aperio CS2 scanner at 40x magnification. Spatial transcriptomics used 10x Genomics Visium and single-cell profiling 10x Genomics Chromium Flex.

Annotations

No image annotations such as tumor outlines are described. Clinical variables come with each patient.

Known limitations

The subset is small, with 10 to 15 patients per cancer type. It is not a download: access needs a project approved by the MOSAIC Window Data Access Committee and a signed agreement, which allows non-commercial academic research only and keeps use within the applying organisation and its country.