Deep Learning for Evolutionary Genomics (DLEG01) – Applications for Evolutionary Biology https://prstats.org/course/deep-learning-for-evolutionary-genomics-dleg01/ Delivered by Dr. Nikolay Oskolkov, a bioinformatician, computational biologist, and data scientist working at the intersection of biology, statistics, and artificial intelligence. His expertise includes machine learning, deep learning, population genomics, ancient DNA, metagenomics, dimensionality reduction, and computational biology. Learn how to apply deep learning and neural networks to complex evolutionary and genomic datasets, with applications to population genomics, molecular evolution, adaptation, introgression, functional genomics, ancient DNA, and evolutionary inference. Modern evolutionary research increasingly generates large and high-dimensional genomic datasets. Deep learning provides powerful approaches for identifying complex patterns in genomic sequences, extracting structure from population genomic data, detecting introgressed regions, analysing ancient DNA, identifying functional genomic elements, and modelling biological sequences. What you'll gain An understanding of the foundations of artificial neural networks and deep learning Practical implementation of deep-learning methods in R and Python Understanding of feed-forward neural networks, CNNs, RNNs, and LSTM networks Applications of neural networks to population and evolutionary genomics Use of neural networks for identifying introgressed genomic regions Autoencoders for dimensionality reduction and representation learning Comparison of autoencoders with PCA, t-SNE, and UMAP Applications of CNNs to gene annotation and functional genomic elements Approaches for analysing DNA as sequence data Deep-learning applications to ancient DNA Introduction to transformer architectures for biological sequence modelling Practical experience extracting biological information from complex, high-dimensional genomic datasets Course format 5-day live, instructor-led online course 25 hours of training Combination of lectures and hands-on practical exercises Practical implementation using R and Python Examples from evolutionary biology, population genomics, functional genomics, metagenomics, and ancient DNA Opportunity to discuss participants' own data All code, datasets, and presentation materials provided All live sessions recorded 30 days of recording access and post-course email support Who is this course for? Evolutionary biologists Evolutionary and population geneticists Evolutionary genomic researchers Molecular evolution researchers Researchers studying adaptation and population history Researchers investigating hybridisation, gene flow, and introgression Ancient DNA and palaeogenomic researchers Functional genomic researchers Bioinformaticians and computational biologists Postgraduate students and early-career researchers Participants should have a basic background in R or Python and some familiarity with biological datasets. Foundational knowledge of statistics, probability, and machine learning is useful but not required, and no previous experience with neural networks or deep learning is necessary. Why take this course? Evolutionary genomics increasingly involves datasets containing millions of genetic variants, complex sequence information, and high-dimensional representations of individuals and populations. Extracting evolutionary information from these datasets presents computational and statistical challenges for which modern deep-learning methods provide an expanding toolkit. For population and evolutionary genomics, neural networks can be used to identify complex genomic patterns associated with processes such as introgression, while autoencoders provide approaches for representation learning and dimensionality reduction. These methods complement established approaches such as PCA, t-SNE, and UMAP and can help reveal structure within high-dimensional genomic datasets. The course also explores applications to functional evolution and ancient DNA. CNNs can be applied to identifying functional genomic elements, while recurrent neural networks and transformer architectures provide approaches for modelling biological sequences. Deep-learning methods for ancient-status inference introduce further applications to evolutionary and palaeogenomic research. Whether you're investigating population history, adaptation, introgression, genomic variation, ancient DNA, functional evolution, or biological sequence evolution, this course provides a practical foundation for incorporating modern deep-learning approaches into evolutionary research. Course dates 5–9 October 2026 25 hours | Live online | £400 Learn more & enrol https://prstats.org/course/deep-learning-for-evolutionary-genomics-dleg01/ Questions? Email: oliver@prstats.org (to subscribe/unsubscribe the EvolDir send mail to evoldir@evoldir.net)