Liftover MCP server

What is Liftover MCP server?

It provides a programmatic interface to the Broad Institute's Liftover tool, enabling genomic coordinate conversion between different genome assemblies (e.g., hg19 to hg38) via the Model Context Protocol. It is intended for developers and bioinformaticians who need to automate liftover operations.

How to use Liftover MCP server?

Install dependencies with uv sync, then run uv run server.py or add the server to your MCP client configuration (e.g., in Cursor) using uv as the command. Invoke the liftover tool with the required from_genome and to_genome parameters plus one of the supported input formats (single position, interval, variant, or HGVS notation).

Key features of Liftover MCP server

• Single position conversion between genome builds
• Interval conversion (start/end coordinates)
• Variant conversion (chr/pos/ref/alt)
• HGVS notation conversion (transcript, c., p.)
• Direct integration with Broad Institute's Liftover web service

Use cases of Liftover MCP server

• Convert a single genomic coordinate from hg19 to hg38
• Map a genomic interval (e.g., 141310715–141310720) to a different assembly
• Transform a variant call (chr, position, ref, alt) between genome builds
• Convert HGVS notation (e.g., c.875A>T) from one reference genome to another

FAQ from Liftover MCP server

What dependencies does Liftover MCP server require?

It requires uv, mcp[cli], fastmcp, beautifulsoup4, and selenium.

Is Liftover MCP server stable?

It is under active development (alpha release v0.1.0-alpha). Features and APIs may change without notice.

Where does the coordinate conversion happen?

All coordinate conversion functionality is provided by the Broad Institute's Liftover web service at liftover.broadinstitute.org.

How do I cite or credit this tool?

Credit the Broad Institute's Liftover service when using this tool in your work. The server itself is licensed under Apache License 2.0.

What input formats are supported?

Single position (chr + position), interval (chr + start + end), variant (chr + position + ref + alt), and HGVS notation (transcript + c_dot + p_dot). The from_genome and to_genome parameters are always required.