VarSome API Update - v 9.4.4

On June 11th 2021, we will update the Stable VarSome API (available at https://stable-api.varsome.com/) from version 9.1.8 to version 9.4.4 in line with our quarterly release procedure.

There are important changes in this release, due to the introduction of the AMP classifier, that may affect your software. We strongly recommend that you follow our published guidelines to increase the speed and reduce the cost of calls to the VarSome API.

We remind you that three versions of the VarSome API are available:

• Live: https://api.varsome.com/ - this is the same as is used by VarSome website
• Stable API & frozen databases: https://stable-api.varsome.com/ (updated quarterly)
• Release candidate: https://staging-api.varsome.com/

Key Functional Changes

AMP
Our new AMP classification can now be enabled by setting the key “add-AMP-annotation=1”. A new ‘AMP_annotation’ result will be returned with a full classification and explanation of the rules and evidence used. This flag also adds the additional databases required for cancer classification.

Please contact us if you would like more information about the data format of the fields and explanations.

ACMG

• Rule PVS1 uses the GnomAD Observed/Expected ratio
• Rules PP2/BP3: Cosmic.FATTHM is used in somatic mode only (not germline)

The GERP conservation scores will no longer be returned, unless you explicitly include them in your add-source-databases, as they are no longer used by the ACMG classifier.

Databases
As part of our AMP implementation, we have added many more databases to VarSome and these will be returned if enable_all is set, or AMP is enabled: 

• NIH GDC
• Pharm GKB
• CBioPortal
• Cancer HotSpots
• ICGC Somatic
• NIH GTex
• Cornell PMKB
• WUSTL CIViC

These new sources also add many additional publication references to the response.

The AMP classifier on VarSome Clinical uses data from JAX CKB. This cannot be licensed for the API.

dbNSFP data format change
The entries from dbNSFP are now arrays providing values for every transcript dbNSFP uses. New array fields called “ensembl_proteinid” and “ensembl_transcriptid” are provided that are formatted like this:

"ensembl_proteinid": [

                "ENSP00000249806",

                "ENSP00000457384",

                "ENSP00000457783",

                "ENSP00000445770",

                "ENSP00000490057",

                "ENSP00000489843"

            ],

            "ensembl_transcriptid": [

                "ENST00000249806",

                "ENST00000565471",

                "ENST00000566347",

                "ENST00000538696",

                "ENST00000637494",

                "ENST00000637667"

            ]

There is a one-to-one relationship between the transcripts and the values provided for each score.

Further Information and Support

Please contact us at the usual address support@varsome.com should you require any additional information or run into any major issues. As ever we look forward to your feedback and suggestions to improve our platform.