Michael Hiller - Computational Biology and Evolutionary Genomics: Discovering phenotype-genotype associations

Previous and current research

Evolution has led to an incredible diversity of phenotypes in all major species clades, exemplified by mammals like bats that fly or dolphins that live entirely in the water. This phenotypic diversity is often the result of changes or loss of ancestral phenotypes or gain of novel phenotypes during evolution. Phenotypic differences between species are due to differences in their genomes. For example, loss of a regulatory element led to the adaptive loss of spines in sticklebacks (PMID 20007865) and the loss of a forebrain enhancer in humans may partially contribute to the gain of human brain complexity (PMID 21390129). The availability of many sequenced genomes together with the availability of hundreds of additional genomes in the near future allows us now to study how molecular and morphological phenotypic diversity is encoded in the genome.

The lab's goal is to develop computational approaches to study phenotype - genotype associations using the power of comparative and evolutionary genomics, followed by experimental verifications.
In particular, we are interested in
What in the genome makes species different at the molecular and morphological level?
How does nature's incredible phenotypic diversity arise?
How does a biological system change and adapt to loss of some of its components?
How does a genome change in evolution as the result of species phenotype changes?

To address these questions, we will develop comparative genomics approaches to associate phenotypic change between species to changes in genomic elements. This will involve exploration-driven approaches that integrate data sources and map functional annotations followed by statistical enrichment tests as well as the development of systematic approaches that aim at predicting specific elements that are associated with a given phenotype of interest. To obtain such predictions of high accuracy, we have to solve problems associated with the incompleteness and inaccuracy of genome assemblies, functional annotations and multiple genome alignments. Subsequent validation experiments in vertebrate, insect or nematode model organisms will test if and how the elements pinpointed by the computational approaches are involved in changing molecular and morphological phenotypes.

Graphical representation of a sequence alignment shows divergence of a gene in an entire phylogeny.

Future prospects and goals

Development of computational approaches to predict associations between phenotype and genotype
Application of these tools to interesting molecular and morphological phenotypic differences in vertebrates, insects and nematodes
Development of novel methods to improve multiple genome alignments

Selected publications