At the Quadram Institute we have a team of bioinformaticians who support our scientists to analyse data on the DNA of microbes. In May they welcomed experts from South Korea and Switzerland to build international partnerships and share bioinformatics expertise.

“We decided to organise a set of workshops at the Quadram Institute to bring together world-leading experts in large-scale analysis of microbial genomes to optimise bioinformatics tools and forge new collaborations,” explains Dr Andrea Telatin, Head of Bioinformatics.

From phylogenetics to protein identification using AI

A key aspect to analysing microbial genomes is making comparisons between the DNA of organisms, whether between different microbes or between bacteria, animals and fungi.

The visiting research group from University of Lausanne in Switzerland are focused on making these comparisons through large-scale genomics across the tree of life. The group led by Dr Natasha Glover and Prof Christophe Dessimoz are responsible for the software ecosystem around a database which keeps up-to-date evolutionary information on almost 3,000 genomes from different organisms and growing, called the Orthologous Matrix (OMA) database.

The University of Lausanne researchers were joined by Professor Martin Steinegger from Seoul National University. His group have developed popular microbial genomics tools, and they apply artificial intelligence to advance structural phylogenetics.

“As the Head of Phylogenomics, I was aware how the two groups’ methodologies can be applied to elucidating bacterial evolution. We were keen to invite them to the Quadram Institute to share expertise,” explains Dr Leo Martins from the Bioinformatics team at Quadram Institute.

Dr Stefano Pascarelli from the University of Lausanne comments on arriving in Norwich, “For me, I was truly impressed by the level of science and support that is found at the Quadram Institute. I observed a great synergy of the research themes in cooperation with an enviable Bioinformatics core facility that explains the growing success within the Norwich Research Park.”

Along with talks, the two-day visit included activities to foster collaboration and dialogue.

“We had a round table discussion with all guests, where we talked about the future of bioinformatic databases, and how we can use them fully for microbial studies” says Leo.

As well as our Bioinformatics team, researchers from across the Quadram Institute joined the workshops to learn and share their latest developments including Dr Dipali Singh who is using mathematical modelling and AI to study the prostate cancer microbiome and Dr Falk Hildebrand whose group investigates the diversity of microorganisms in the gut microbiome through metagenomics.

Dipali says, “Having world-leading experts in genome analysis get together at the Quadram Institute was an incredible opportunity. I gained valuable insights and learned about new tools that will be useful for my research. I am excited about the prospect of collaborating with these experts in the future.”

Falk adds, “It was a fantastic opportunity to welcome some of the forerunners of bioinformatic research to the Quadram Institute. I enjoyed spending time with these researchers, some of whom’s software I have been using for years. I learnt more about the development process and reasoning in these hugely influential software. Going forward I think our interactions and exchanges will lead to future collaborations. ”

Improving and extending the use of existing bioinformatic tools

“It was a very successful activity. At the same time as the guests presented state-of-the-art methods for the kind of bioinformatic analysis we do at the institute, we provided guests with specific needs and challenges posed by bacterial genomes,” reflects Leo.

Professor Martin Steinegger says, “Assigning functions labels to metagenomic open reading frames is one of the major challenges in advancing microbial research. With the advances of next-generation structure predictors like AlphaFold2 and extensive databases structural database, we have a unique opportunity to improve annotation.”

Dr Irene Julca Chavez from the University of Lausanne highlights the benefits from both sides, “I believe this was a very good initiative. Currently, I work in comparative genomics, and our lab has many tools that assist with analysis in this field. However, the discussions with experts from different fields during the meetings at the Quadram Institute made us aware of the lack of resources for certain taxonomic groups, such as bacteria and viruses”.

Dr Dave Moi, from the University of Lausanne continues, “With the new tools that are emerging today, we can start making progress on redefining the viral taxonomy, annotating proteins of unknown function and resolve difficult phylogenies of highly diverged protein families. Going forward at the Dessimoz lab, it looks like we’re going to be integrating more and more of the structural perspective into our workflows and tools.”

Future collaborations

Along with discussions and work to develop existing bioinformatics tools the international workshop sparked ideas for future collaborations.

“I think this experience will help us expand some lines of research in our group and promote new collaborations,” highlights Irene.

Stefano continues, “Our trip to Norwich led to a fantastic exchange of ideas and knowledge that will possibly lead to exciting new developments. We convened towards the need of shifting the attention of our research towards scalable methods that employ the additional layer of information coming from protein 3D structures.”

Leo from the Quadram Institute Bioinformatics team adds, “There are plenty of opportunities for future collaborations, both in terms of method and workflow development but also in terms of practical applications to our increasingly large data sets generated here.”

Dr Natasha Glover from the University of Lausanne says, “There is significant potential for collaboration between our group and the Quadram Institute, particularly in developing a database and tools for tracking prokaryote-specific evolution. The future of genomic analysis looks promising, with recent advancements in structural genomics now enabling us to recognise homologous genes where sequence similarity fails.”

Martin adds, “We see multiple ways to collaborate with researchers at the Quadram Institute to enhance metagenomic analysis using structural aligners like Foldseek and Foldseek-multimer, implement our metagenomic classifier Metabuli well as the structure predictor ColabFold.”

Natasha Glover concludes, “This experience has reaffirmed the necessity and benefits of international exchanges in pushing the boundaries of genomics.”

The workshop was funded through BBSRC’s International Partnership Awards