# Migration Guide

This page describes various changes deployers will need to make as they deploy newer versions of the portal.

# v4 -> v5

• All fusion profiles are now required to be migrated to structural variant format. One can use this migration tool to migrate the fusion files.
• All fusion files on datahub were migrated to the structural variant format and their molecular profile ids were renamed from {study_id}_fusion to {study_id}_structural_variants. If you are using these datahub files one would need to re-import them.
• Study view user setting will be outdated after migration, please follow Clear Study View User settings section in Session Service Management
• The new default set of transcripts for each gene has changed from uniprot to mskcc. See the Mutation Data Annotation Section for more details. To keep the old set of default transcripts add the property genomenexus.isoform_override_source=uniprot to the properties file.

See the v5.0.0 release notes for more details.

# v3 -> v4

• Introduces logback package for logging. If you don't have any custom log4j.properties file, no changes are necessary
• Cleans up several old databases (PR). In theory the migration should be seamless, since the docker container detects an old database version and migrates it automatically.

See the v4.0.0 release notes for more details.

# v2 -> v3

• Session service is now required to be set up. You can't run it without session service. The recommended way to run cBioPortal is to use the Docker Compose instructions.

# v1 -> v2

• Changes cBioPortal to a Single Page App (SPA) written in React, Mobx and bootstrap that uses a REST API. It shouldn't change anything for a deployer.