MAGI

MAGI (Mutation Annotation and Genome Interpretation) is an open-source, web-based platform for interactive visualization, integration, and collaborative annotation of cancer genomic aberrations across public and private datasets. It is designed to enable bidirectional analysis: investigators can contextualize their own tumor profiles using large public cohorts while also contributing expert annotations that enrich public resources.

MAGI provides linked, interactive views over multiple data types for user-selected genes and samples, including single-nucleotide variants and small indels, copy-number aberrations, gene expression, and protein–protein interaction/contextual annotations, with real-time zooming, panning, and filtering. The system comes preloaded with TCGA Pan-Cancer data and associated sample attributes (e.g., survival time, gender, tumor purity estimates), and integrates external protein domain and interaction annotations. Users can upload private datasets (mutations, expression, methylation, and sample metadata) via a web form without local installation; MAGI then automatically generates dataset summaries with interactive plots, sortable tables, and pathway analysis.

A core feature is its collaborative annotation layer, supporting literature citations, text comments, and voting for both public and user-provided events; the interface is initialized with tens of thousands of curated protein sequence change annotations from sources such as the Database of Curated Mutations (DoCM) and automated literature searches. MAGI also includes per-sample “n = 1” views that aggregate aberrations and their annotations, a bookmarking mechanism for sharing interactive views (optionally including private data), export of publication-quality graphics, and an interactive compute engine for statistical association testing between genomic events and sample attributes across combined datasets.