LAMP-LD – Local Ancestry Inference in Admixed populations

Software package for local ancestry inference in admixed populations. It takes as input genotype data from admixed samples and produces a local ancestry (chromosome painting) according to the ancestry at every locus in the genome. Several versions are included in the software package:

WINPOP: window-based algorithm that use an improved modeling of the recombinations events within each window. This leads to a better accuracy for recently admixed populations where the ancestral populations are closely related (e.g., European Americans or Japanese-Chinese, etc.), in case the ancestral populations are known. WINPOP does not model haplotypes and assumes SNPs to be independent conditional on their ancestry (Pasaniuc et al. 2009).

LAMP-LD: window-based algorithm combined within a hierarchical Hidden Markov Model to represent haplotypes in the population. It can process 2,3 or 5 mixing populations, and its short per-sample processing time makes it suitable for analyzing large datasets of dense SNP panels (Baren et al. 2012).

LAMP-HAP: similar to LAMP-LD but processes haplotype data when high-quality phasing is available (e.g. in families), and utilizes trio nuclear family designs to improve estimation accuracy.

Download LAMP-LD version 1.1 here.


Baran, Yael, Bogdan Pasaniuc, Sriram Sankararaman, Dara G. Torgerson, Christopher Gignoux, Celeste Eng, William Rodriguez-Cintron, et al. 2012. Fast and accurate inference of local ancestry in Latino populations. Bioinformatics 28, no. 10 (April 11): 1359–1367.
Pasaniuc, B., S. Sankararaman, G. Kimmel, and E. Halperin. 2009. Inference of locus-specific ancestry in closely related populations. Bioinformatics 25, no. 12 (May 28): i213–i221.