Periodic Reporting for period 3 - DiversiPHI (Predicting the evolution of complex phage-host interactions)
Reporting period: 2023-04-01 to 2024-09-30
What is the problem/issue being addressed?
Bacteriophages are viruses that infect bacteria, and like all viruses they require their host to replicate. Both phages and their hosts are evolving entities, and the outcome of their co-evolutionary process is an open question that depends on many factors. DiversiPHI aims to understand how these many factors are related and integrated, by asking:
(1) Can a phage evolve to infect a given bacterium?
(2) What phage and host characteristics are important for this ability?
(3) Can phage-host compatibility after short-term evolution be computationally predicted?
To address these questions, we will use complementary computational and experimental approaches. First, we will use big data analysis to measure a range of phage and host factors that may be relevant to the interaction (genetic, functional, and ecological factors). Second, these factors will be combined by using artificial intelligence techniques with the aim of figuring out how and why they are important in the infection process. Third, we will use experimental evolution to test host-range evolution in a laboratory setting.
Why is it important for society?
Bacteriophages have important applications in e.g. medicine and biotechnology, and far-ranging effects on microbial community functioning in all environments. In turn, microbial communities have important roles in human, plant, and animal health and in a range of industrial processes. Thus, it is very important to be able to predict which bacteria might be (more readily) targeted by a certain phage, to understand the impact of phages on different bacteria, and by extension their effect on complex microbial communities. DiversiPHI not only aims to deliver such predictions, but also to shed light on the mechanisms underlying these interactions.
What are the overall objectives?
The objectives of the DiversiPHI project are to:
(1) collect genomic and metagenomic data for a wide range of bacteriophages that are present in public data repositories from a range of different (micro-) biomes;
(2) create computational analysis methods to measure genomic, functional, and ecological features, known as “phage-host interaction” (PHI) factors;
(3) integrate the PHI factors into a phage-host predictor and to compare the prediction accuracy for different training and testing datasets;
(4) identify which PHI factors are important for predicting bacteriophage host-range;
(5) to enhance our understanding of bacteriophage-bacterial co-evolution across biomes.
The outlook of the project is to understand the role of viruses in structuring the microbial biosphere, while generating critical insights and innovative computational tools to further the viral ecology field.
Bacteriophages are viruses that infect bacteria, and like all viruses they require their host to replicate. Both phages and their hosts are evolving entities, and the outcome of their co-evolutionary process is an open question that depends on many factors. DiversiPHI aims to understand how these many factors are related and integrated, by asking:
(1) Can a phage evolve to infect a given bacterium?
(2) What phage and host characteristics are important for this ability?
(3) Can phage-host compatibility after short-term evolution be computationally predicted?
To address these questions, we will use complementary computational and experimental approaches. First, we will use big data analysis to measure a range of phage and host factors that may be relevant to the interaction (genetic, functional, and ecological factors). Second, these factors will be combined by using artificial intelligence techniques with the aim of figuring out how and why they are important in the infection process. Third, we will use experimental evolution to test host-range evolution in a laboratory setting.
Why is it important for society?
Bacteriophages have important applications in e.g. medicine and biotechnology, and far-ranging effects on microbial community functioning in all environments. In turn, microbial communities have important roles in human, plant, and animal health and in a range of industrial processes. Thus, it is very important to be able to predict which bacteria might be (more readily) targeted by a certain phage, to understand the impact of phages on different bacteria, and by extension their effect on complex microbial communities. DiversiPHI not only aims to deliver such predictions, but also to shed light on the mechanisms underlying these interactions.
What are the overall objectives?
The objectives of the DiversiPHI project are to:
(1) collect genomic and metagenomic data for a wide range of bacteriophages that are present in public data repositories from a range of different (micro-) biomes;
(2) create computational analysis methods to measure genomic, functional, and ecological features, known as “phage-host interaction” (PHI) factors;
(3) integrate the PHI factors into a phage-host predictor and to compare the prediction accuracy for different training and testing datasets;
(4) identify which PHI factors are important for predicting bacteriophage host-range;
(5) to enhance our understanding of bacteriophage-bacterial co-evolution across biomes.
The outlook of the project is to understand the role of viruses in structuring the microbial biosphere, while generating critical insights and innovative computational tools to further the viral ecology field.
We have by started off the DiversiPHI project by developing tools and data mining studies to extract important ecological, functional, and taxonomic PHI factors from genomic and metagenomic datasets.
(1) Ecology: our definition of the “social niche breadth” (SNB) score allowed us to quantify microbial niche space directly from metagenomic data (preprint: https://www.biorxiv.org/content/10.1101/2022.07.21.500953v1(opens in new window)).
(2) Ecology: we have identified viral abundance and diversity as major ecological factors that drive the selection and maintenance of CRISPR-Cas in microbial ecosystems (Meaden et al., Curr Biol 2022).
(3) Functions: we have built a proof-of-concept predictor for bacteriophage protein functions based on machine learning (Pappas & Dutilh, BMC Bioinf 2022).
(4) Taxonomy: we have contributed to the establishment of novel viral taxonomy, including the Crassvirales order (https://ictv.global/taxonomy/taxondetails?taxnode_id=202113057(opens in new window)).
We have also organized the first hands-on “Viromics Workshop” in Jena in February 2022, which attracted 35 students, mostly from Europe. While ~20 students came live to Jena, the course was available in a hybrid format which also allowed international students to attend. We discussed topics including virus discovery in metagenome data, different virus discovery tools, protein clustering and annotation, taxonomic classification, and ecological dynamics. The course material is available open access (https://mgxlab.github.io/Viromics-Workshop-MGX/(opens in new window)).
(1) Ecology: our definition of the “social niche breadth” (SNB) score allowed us to quantify microbial niche space directly from metagenomic data (preprint: https://www.biorxiv.org/content/10.1101/2022.07.21.500953v1(opens in new window)).
(2) Ecology: we have identified viral abundance and diversity as major ecological factors that drive the selection and maintenance of CRISPR-Cas in microbial ecosystems (Meaden et al., Curr Biol 2022).
(3) Functions: we have built a proof-of-concept predictor for bacteriophage protein functions based on machine learning (Pappas & Dutilh, BMC Bioinf 2022).
(4) Taxonomy: we have contributed to the establishment of novel viral taxonomy, including the Crassvirales order (https://ictv.global/taxonomy/taxondetails?taxnode_id=202113057(opens in new window)).
We have also organized the first hands-on “Viromics Workshop” in Jena in February 2022, which attracted 35 students, mostly from Europe. While ~20 students came live to Jena, the course was available in a hybrid format which also allowed international students to attend. We discussed topics including virus discovery in metagenome data, different virus discovery tools, protein clustering and annotation, taxonomic classification, and ecological dynamics. The course material is available open access (https://mgxlab.github.io/Viromics-Workshop-MGX/(opens in new window)).
A big change of the DiversiPHI project was the move of the PI to a new host institution (Friedrich Schiller University Jena), where an integrated dry-/wet-lab was established. The move has now been finalized, and the remainder of the project can now be carried out there.
Two new PhD students have started in Jena in September 2022:
(1) the first will collaborate with the Wet Lab Manager and other group members to develop DiversiPHI's host switching experiments;
(2) the second will work on integrating diverse data including PHI factors using machine learning.
These new project members will work on the remaining objectives of the DiversiPHI project in the coming years.
Two new PhD students have started in Jena in September 2022:
(1) the first will collaborate with the Wet Lab Manager and other group members to develop DiversiPHI's host switching experiments;
(2) the second will work on integrating diverse data including PHI factors using machine learning.
These new project members will work on the remaining objectives of the DiversiPHI project in the coming years.