Dr Hajk-Georg Drost

My academic path was enabled by studying computer science and bioinformatics with a strong focus on statistical learning, machine learning, and predictive modeling with applications in comparative genomics and evolutionary developmental biology. I obtained my BSc and MSc in Bioinformatics, and a PhD in Computer Science at the Institute of Computer Science – Martin-Luther University Halle, Germany where I studied natural variation and machine learning through the lens of the developmental hourglass model with Ivo Grosse and Marcel Quint. New questions concerning the dynamics and epigenetic control of genomes led me to pursue a postdoc in the lab of Jerzy Paszkowski at the Sainsbury Laboratory and Genetics Department of the University of Cambridge, where I studied the epigenetic control of transposable elements and how these elements can generate natural variation in genomic landscapes. As a next step, I wanted to integrate my insights from evo-devo and (epi)-genomics research and joined the second lab of Elliot Meyerowitz (first lab at Caltech) as a senior postdoc at the Sainsbury Laboratory in Cambridge. With Elliot, I sought to understand how regulatory changes differ across organ evolution in plants and animals to unveil the molecular processes whereby new organ morphology is integrated across diverse plant and animal species. Here, I was especially interested in the roles of orthologous genes, lncRNAs and circRNAs in regulating these developmental processes. At Cambridge, I fortunate to be elected Fellow of the Cambridge Philosophical Society and Postdoctoral Affiliate of Trinity College.

After studying the mechanisms generating natural variation on different levels of organismal complexity, I was offered to start a Computational Biology Group in the Department of Molecular Biology (led by Detlef Weigel) at the Max Planck Institute for Biology Tubingen, Germany. There I contributed to the growing field of Machine Learning in biology and life sciences. My team explored how causal hypotheses can be tested when inferring gene regulatory interactions from multi-omics datasets to understand how biological function is retained evolutionary time or diversifies with perturbation effects that can lead to (developmental) diseases (e.g. cancer). For this attempt, I brought together a wide range of experts from computer science, bioinformatics, and molecular biology backgrounds to develop the tools and methodologies to approach this question.

My team’s research led to breakthrough technologies and discoveries at tree-of-life scale which allowed me to secure a Associate Professorship in Digital Biology within the Division of Computational Biology at the School of Life Sciences (University of Dundee), Scotland, UK. Here, we seek to make long-term investments into the intelligent software architectures and methodological innovation in Generative (Life Science) AI that is required to translate digital biology research and machine learning methodologies into bio-pharmaceutical and medical application. In Dundee (Scotland), I was fortunate to be awarded the Royal Society Wolfson Fellowship which supports our efforts to harness machine learning to uncover fundamental principles of developmental regulation to guide drug discovery efforts to prevent or mitigate the effects of human developmental diseases.