Periodic Reporting for period 1 - Comp-Biodiv-GR (Strengthening Computational Biodiversity Research in Greece)
Reporting period: 2023-01-01 to 2024-03-31
Greece is a biodiversity hotspot due to its complex geology and climate variation, with a high percentage of unique species, particularly on islands like Crete. Preserving this biodiversity is a key focus for Crete's regional government, the national government of Greece, and the EU. Bioinformatics, the analysis of biological data using computer software, is crucial in biodiversity research, but software development for biodiversity data analysis is underdeveloped in Greece. Additionally, the country has faced significant brain drain due to the financial crisis, with many biologists and computer scientists leaving for PhDs abroad and not returning. Few international scientists move to Greece, further exacerbating the issue.
Our project aims to reverse brain drain and foster computational biodiversity research in Greece by establishing a visible research group at ICS-FORTH, Greece’s leading computer science institute. This group will focus on developing open-source bioinformatics tools and methods for biodiversity data analysis, a vital first step in biodiversity preservation. We work interdisciplinarily with regional institutions conducting fieldwork and data analyses.
Our research enhances biodiversity protection and awareness in Greece, the EU, and beyond. It fosters brain gain in Europe's periphery and trains young researchers, contributing to the knowledge-based society in academia and industry. The project aims to maintain Europe’s leading role in biodiversity research and strengthen Greece as both a physical and computational biodiversity hotspot.
Our project aims to reverse brain drain and foster computational biodiversity research in Greece by establishing a visible research group at ICS-FORTH, Greece’s leading computer science institute. This group will focus on developing open-source bioinformatics tools and methods for biodiversity data analysis, a vital first step in biodiversity preservation. We work interdisciplinarily with regional institutions conducting fieldwork and data analyses.
Our research enhances biodiversity protection and awareness in Greece, the EU, and beyond. It fosters brain gain in Europe's periphery and trains young researchers, contributing to the knowledge-based society in academia and industry. The project aims to maintain Europe’s leading role in biodiversity research and strengthen Greece as both a physical and computational biodiversity hotspot.
For up-to-date and more comprehensive information, please visit our project web-site at www.biocomp.gr and in particular, the news, talks, and publications sections that are updated frequently.
In the first 15 months, we have:
- Hired an international research team with members from Germany, Austria, Spain, the US, and Greece.
- Arranged for a visiting professor from the US to stay in Crete for five months in early 2025.
- Contributed to training future researchers by organizing a summer school on computational molecular evolution, a workshop on Biodiversity informatics in Crete, and teaching a Master-level course via Zoom
with the University of Crete and Karlsruhe Institute of Technology in Germany.
- Established a new seminar on “Reproducibility in Bioinformatics” at the University of Crete.
- Given over 10 scientific talks at various research institutions.
- Conducted public outreach activities, visiting four primary schools in Crete to encourage interest in STEM and biodiversity.
We have published or are about to publish 12 research papers that typically describe the open-source Bioinformatics tools which we make freely available to the scientific community for generating novel knowledge (see https://www.biocomp.gr/publications.html(opens in new window)). A notable tool with potential broad impact is ecofreq which allows to scale computer processor power consumption, depending on the fraction of renewable energy in the overall energy mix (i.e. processor will run at a lower speed when few renewables contribute to the energy mix and vice versa). This allows for reducing both, the cost of energy required for computing as well as the CO2 footprint of computations. Its application extends far beyond Bioinformatics to any type of compute-intensive operations in academia and industry.
To prevent brain drain in a sustainable manner in Greece, we have drafted and proposed two hands-on and realistic institutional reforms as well as one visionary reform. The visionary reform foresees a dedicated female researchers and professors program (such programs already exist in numerous other EU countries) with the aim to strengthen female representation in Greek academic governance bodies where women are currently dramatically underrepresented.
In the first 15 months, we have:
- Hired an international research team with members from Germany, Austria, Spain, the US, and Greece.
- Arranged for a visiting professor from the US to stay in Crete for five months in early 2025.
- Contributed to training future researchers by organizing a summer school on computational molecular evolution, a workshop on Biodiversity informatics in Crete, and teaching a Master-level course via Zoom
with the University of Crete and Karlsruhe Institute of Technology in Germany.
- Established a new seminar on “Reproducibility in Bioinformatics” at the University of Crete.
- Given over 10 scientific talks at various research institutions.
- Conducted public outreach activities, visiting four primary schools in Crete to encourage interest in STEM and biodiversity.
We have published or are about to publish 12 research papers that typically describe the open-source Bioinformatics tools which we make freely available to the scientific community for generating novel knowledge (see https://www.biocomp.gr/publications.html(opens in new window)). A notable tool with potential broad impact is ecofreq which allows to scale computer processor power consumption, depending on the fraction of renewable energy in the overall energy mix (i.e. processor will run at a lower speed when few renewables contribute to the energy mix and vice versa). This allows for reducing both, the cost of energy required for computing as well as the CO2 footprint of computations. Its application extends far beyond Bioinformatics to any type of compute-intensive operations in academia and industry.
To prevent brain drain in a sustainable manner in Greece, we have drafted and proposed two hands-on and realistic institutional reforms as well as one visionary reform. The visionary reform foresees a dedicated female researchers and professors program (such programs already exist in numerous other EU countries) with the aim to strengthen female representation in Greek academic governance bodies where women are currently dramatically underrepresented.
Progress beyond the stat-of-the art has been achieved by the plethora of original scientific publications (and accompanying open-source software tools) that have already been published in the first 15 months. Apart from the aforementioned work on energy efficiency, or more specifically minimizing the cost and CO2 footprint of computations, we have focused on deploying machine learning methods in novel ways to accelerate biodiversity data analyses, but also to highlight weaknesses of current scientific approaches. We have also opened a new line of research in the area of computational linguistics and more specifically on reconstructing the evolutionary history of natural languages.
Numerous of the run time improvements that we have achieved either via machine learning, parallel computing, or algorithmic research will have immediate impact as these are improvements to existing scientific software tools that are already being widely used, for instance RAxML-NG (2477 citations based on Google Scholar - respective paper describing the software available at: https://academic.oup.com/bioinformatics/article/35/21/4453/5487384 ).
Research-wise we expect to generate, for instance, novel methods for analyzing biodiversity data from so-called Malaise traps that are used for monitoring the highly important insect biodiversity, develop machine-learning based methods to quantify, a priori (before actually running the analysis), the expected uncertainty of analyzing given DNA input data under specific models and/or algorithms, develop machine learning approaches to automatically extract morphological features of insects from respective photographs, contribute to analyzing and assembling the Cretan pan-genome (whole genome data from 50-100 individuals from Crete), developing novel tools in the area of biogeography and ancient DNA data analysis etc. In general, we will strive to attain a balance between pure computational method development and analyzing real biological data that will mainly be collected in Crete.
The aforementioned teaching activities will be continued throughout the duration of the project and we will certainly organize additional public outreach events.
Finally, the potential commercialization of our ecofreq tool and its deployment in compute-intensive industry sectors will be one key priority.
Numerous of the run time improvements that we have achieved either via machine learning, parallel computing, or algorithmic research will have immediate impact as these are improvements to existing scientific software tools that are already being widely used, for instance RAxML-NG (2477 citations based on Google Scholar - respective paper describing the software available at: https://academic.oup.com/bioinformatics/article/35/21/4453/5487384 ).
Research-wise we expect to generate, for instance, novel methods for analyzing biodiversity data from so-called Malaise traps that are used for monitoring the highly important insect biodiversity, develop machine-learning based methods to quantify, a priori (before actually running the analysis), the expected uncertainty of analyzing given DNA input data under specific models and/or algorithms, develop machine learning approaches to automatically extract morphological features of insects from respective photographs, contribute to analyzing and assembling the Cretan pan-genome (whole genome data from 50-100 individuals from Crete), developing novel tools in the area of biogeography and ancient DNA data analysis etc. In general, we will strive to attain a balance between pure computational method development and analyzing real biological data that will mainly be collected in Crete.
The aforementioned teaching activities will be continued throughout the duration of the project and we will certainly organize additional public outreach events.
Finally, the potential commercialization of our ecofreq tool and its deployment in compute-intensive industry sectors will be one key priority.