Lab logo
CompStatGen Lab
FIMM · Helsinki

What we study

Research

We focus on two interconnected research areas: developing scalable computational methods for genomic data, and applying them to biobanks to understand the genetic basis of human disease.

Structural variation research

Theme 01

Structural Variation & Disease

Large genomic rearrangements, such as deletions, duplications and inversions are often overlooked in standard GWAS studies, yet account for a substantial fraction of heritable disease risk. We develop methods to detect, phase, and functionally interpret structural variants at biobank scale, integrating them with transcriptomic and proteomic data.

SV detection Multi-omics UK Biobank FinnGen
Population haplotypes research

Theme 02

Population Haplotypes & Identity-by-Descent

We analyze genetic relationships within and between populations by identifying shared haplotype segments: regions inherited from a recent common ancestor. This illuminates human population history, enables powerful disease mapping through IBD, and underpins accurate genotype imputation.

Haplotype phasing IBD Population genetics SHAPEIT5
Low coverage sequencing research

Theme 03

Low-Coverage Sequencing & Imputation

SNP arrays have dominated human genetics for two decades, but low-coverage whole-genome sequencing is rapidly becoming a cost-effective alternative. We develop statistical methods such as GLIMPSE that recover accurate genotypes, including rare variants, from sequencing depths as low as 0.1×.

lcWGS GLIMPSE2 Rare variants Ancient DNA
Computational genomics algorithms

Theme 04

Scalable Computational Algorithms

All our biological questions are ultimately constrained by computation. We design algorithms that scale to millions of genomes. Core methodological themes include Hidden Markov models on haplotype space, Positional Burrows-Wheeler Transform (PBWT), extensions of the Li–Stephens model, and compressed reference panel representations.

HMM PBWT Li-Stephens IMPUTE5