Liftover

GWASLab can directly liftover the positions of variants in sumstats using a fast chain-based implementation.

.liftover()

mysumstats.liftover(from_build="19", 
                    to_build="38",
                    remove=True)

Note

GWASLab will only liftover basepair positions. If needed, please perform harmonization using the reference file of the target build.

Implementation Details

The liftover implementation uses the sumstats_liftover package, which directly parses UCSC chain files and performs vectorized coordinate conversion for optimal performance.

Key features: - Fast chain-based conversion: Directly parses UCSC chain files with vectorized operations - Built-in chain files: Automatically uses built-in chain files for hg19↔hg38 conversions (downloaded from UCSC if not available) - Chromosome mismatch detection: Variants mapped to different chromosomes are automatically detected and treated as unmapped

Options

`.liftover()` options	DataType	Description	Default
`from_build`	`"19"` or `"38"`	Original genome build	-
`to_build`	`"19"` or `"38"`	Target genome build	-
`chain_path`	`str`	Path to UCSC chain file (optional). If provided, `from_build` and `to_build` are optional	`None`
`remove`	`boolean`	If True, remove unmapped variants (including chromosome mismatches)	`True`

Quote

This method uses the sumstats_liftover package, which provides fast chain-based coordinate conversion using vectorized operations.

Example

# Basic usage with build numbers
mysumstats.liftover(from_build="19", to_build="38")

# Using a custom chain file
mysumstats.liftover(chain_path="/path/to/chain/file.chain")

# Keep unmapped variants in the output
mysumstats.liftover(from_build="19", to_build="38", remove=False)