AlmaasLab NTNU

Research

The AlmaasLab works in systems biology: understanding the function of systems of biological components, where the interactions — and the webs they generate — matter as much as the parts. Computational and theoretical analyses go hand in hand with our own experiments.

Complex network analysis

From biological to social and technological networks: we develop and apply methods for understanding the structure and dynamics of interaction webs.

While much of our activity is in biological networks, we are interested in a wide range of networks — from social to technological ones. We develop methods for network analysis, including tools for gene co-expression networks (wTO, CoDiNA, csdR) and decompositions of weighted networks such as the s-core. We apply them to large-scale health data: phenome-wide association networks and genome-wide association studies in the HUNT study, and gene expression network studies of diseases including breast cancer, bipolar disorder, multiple sclerosis and Alzheimer’s disease. By aligning the metal-binding sites of all metalloproteins in the Protein Data Bank — more than 23,000 sites — we have built a similarity network of binding environments that recapitulates metal coordination chemistry and enzyme function, and reveals recurrent binding-site geometries shared by proteins with divergent sequences, including unexpected drug off-targets. Where network analysis is not obviously applicable, we let network thinking inspire our approaches.

Genome-scale metabolism

We build and analyse genome-scale metabolic models for a range of organisms, calibrated against our own growth and biomass-composition experiments.

Computational modeling of whole-cell metabolism has become a reality. We build genome-scale metabolic reconstructions for a variety of organisms — from E. coli, Bacillus subtilis and Streptomyces coelicolor to yeasts and the oleaginous Aurantiochytrium — and develop the methods around them: enzyme- and temperature-constrained models, dynamic flux balance analysis of phenomena such as diauxic growth, handling of model uncertainty, and prediction of strain-engineering strategies for improved production. We have shown that biomass composition changes with environmental conditions, and we measure it experimentally in our own lab, together with growth properties, to calibrate the models. Our software — ModelExplorer, ErrorTracer, and pipelines that automatically assemble draft reconstructions from KEGG or from biosynthetic gene clusters — makes model construction reliable. The models can also be turned around: instead of predicting a phenotype from an environment, we use them to design selection niches — growth environments in which evolution itself is steered toward a desired metabolic phenotype. Wine yeasts adaptively evolved in such model-designed niches acquired the metabolic phenotypes the models had predicted, turning laboratory evolution from something we observe into something we can program.

Epidemiology on networks

Disease spread on human contact networks — from antibiotic resistance in care facilities to household-targeted COVID-19 testing and vaccination strategies.

The spread of disease has radically changed human society multiple times through history. We use complex network theory and agent-based modeling to study the interplay between disease dynamics and the structure of human contact networks. Our individual-based framework simulates transmission on temporal contact networks at high resolution — within primary schools, hospitals and care facilities, and across Norway’s network of municipalities. With it, we have shown that targeting whole households rather than individuals makes both testing and vaccination strategies markedly more effective at containing a pandemic — findings we carried into the national public debate during COVID-19. We have also modeled the spread of antibiotic resistance in bacterial metapopulations and the transmission of multi-resistant gonorrhea through contact networks in Norway.

Phage–bacteria evolution

Experimental evolution of bacteriophage–bacteria interactions, exploring phages as an alternative to antibiotics.

The rapid rise of microbes resistant to single or multiple antibiotics is a major challenge to modern medicine, and phages — viruses that uniquely target bacteria — provide an alternative approach for treating bacterial infections. We conduct experimental evolution in designer systems with specific phage–bacteria interactions, following how bacterial resistance against phages emerges under different growth environments and phage pressures. To characterize these interactions, we have developed our own fluorescence-imaging platform — hardware, software and image analysis built in-house — that follows phage plaques on bacterial lawns as they form and grow, paired with reaction–diffusion models of plaque development. We have also automated the Appelmans protocol for expanding the host range of phage cocktails, and demonstrated in Atlantic salmon rearing systems that phage therapy can prevent infection while leaving the surrounding water microbiota essentially undisturbed.

Collaboration

Our research is carried out in close collaboration with experimental and clinical groups, through national initiatives such as the Centre for Digital Life Norway, ELIXIR Norway and the HUNT study, and in European consortia under ERA-NET and Horizon programmes. Research projects the lab has taken part in include: