Research
Our group develops novel computational methods that combine big data algorithms and machine learning to gain insights into unexplored microbial communities.
Methods
MMseqs2 (Many-against-Many sequence searching) is a software suite to search and cluster huge protein and nucleotide sequence sets. MMseqs2 can run 10000 times faster than BLAST. It can perform profile searches with the same sensitivity as PSI-BLAST at over 400 times its speed.
ColabFold is an easy-to-use environment for fast and convenient protein structure predictions. Its structure prediction is powered by AlphaFold2 and RoseTTAFold combined with a fast multiple sequence alignment generation stage using MMseqs2, which speeds up the MSA generation by a factor of 16 over the AlphaFold system.
Foldseek is a software suite for searching and clustering protein structures. It is 600,000 times faster than the fastest state-of-the-art aligners. Allowing to query millions of structures in seconds.
Linclust is a method that can cluster sequences down to 50% pairwise sequence similarity and its runtime scales linearly with the input set size, not quadratically as in conventional algorithms. It is >1000 times faster compared to its competitors.
Plass is a software to assemble short read sequencing data on a protein level. The main purpose of Plass is the assembly of complex metagenomic datasets. It assembles 10 times more protein residues in soil metagenomes than Megahit.
Conterminator is an efficient method for detecting incorrectly labeled sequences across kingdoms by an exhaustive all-against-all sequence comparison.