As modern farming becomes more intensive and threatens the survival of various species, Dr. Dirk Steinke’s research team is developing new ways to monitor insect populations in real time. They used machine learning and genetic tools to create efficient, automated methods for measuring biodiversity and assessing environmental impacts. These innovations help guide habitat management at both local and broader scales.
As global food demand increases, more land is converted to farming. This intensive agriculture leads to habitat destruction, pollution and disruption of natural processes, affecting all living things, especially small creatures. Arthropods (insects, spiders and their relatives) represent a huge portion of Earth’s species and perform many vital functions in nature. However, we do not fully understand the extent of arthropod diversity or how modern farming and environmental stresses affect these insects. We need continuous monitoring of insect communities to improve conservation efforts and develop farming practices that protect both biodiversity and food production.
Working with the Centre for Biodiversity Genomics, Dr. Dirk Steinke’s team established methods to measure insect biodiversity and developed practical ways to monitor it in real time. They used new computer learning techniques to tackle the challenge of classifying arthropods. The project focused on handling complex biological data from genetic sequences, visual features and written descriptions. By improving automated classification and prediction accuracy, they created methods that can be expanded and adapted to study other forms of biological data.
The research produced several innovations in automating biological data analysis. The team developed tools that can process and interpret large amounts of diverse biological information efficiently. They found that total RNA sequencing (a method of studying genetic material) works better than other techniques for understanding microscopic organism diversity, especially in complex organisms. They also showed that specific genetic markers (16S rRNA) can reveal environmental stress by analyzing changes in microbial communities. The team also created a computer program that can accurately pull arthropod species names from scientific literature, improving the pace of species documentation. The team combined DNA analysis, direct observation and chemical analysis for case studies of specific groups, such as invasive species like green crabs. Furthermore, the team developed a low-cost, high-throughput imaging system capable of digitizing up to 200 insect specimens per hour, dramatically increasing the efficiency of insect data collection.
Steinke’s research team provided practical value by improving biodiversity monitoring through cost-effective methods. Using computer learning to analyze large sets of genetic, photographic and ecological data, his team gives a clearer picture of insect diversity in farming ecosystems. This helps evaluate harmful farming practices and assess whether conservation efforts are working. The research goes beyond just identifying species by studying their roles in the ecosystem including their diet and ecological functions, which helps guide habitat management, pest control and pollinator protection. By combining various scientific methods, the project shows how genetic differences, diet and gender affect species’ roles in nature and provides better ways to assess biodiversity and make informed decisions.
Cordone, G., Lozada, M., Vilacoba, E., Thalinger, B., Bigatti, G., Lijtmaer, D. A., Steinke, D., and Galván, D. E. (2022). Metabarcoding, direct stomach observation and stable isotope analysis reveal a highly diverse diet for the invasive green crab in Atlantic Patagonia. Biological Invasions, 24(2), 505–526. https://doi.org/10.1007/s10530-021-02659-5
Hempel, C. A., Buchner, D., Mack, L., Brasseur, M. V., Tulpan, D., Leese, F., and Steinke, D. (2023). Predicting environmental stressor levels with machine learning: A comparison between amplicon sequencing, metagenomics and total RNA sequencing based on taxonomically assigned data. Frontiers in Microbiology, 14, 1217750. https://doi.org/10.3389/fmicb.2023.1217750
Hempel, C. A., Carson, S. E. E., Elliott, T. A., Adamowicz, S. J., and Steinke, D. (2023). Reconstruction of small subunit ribosomal RNA from high‐throughput sequencing data: A comparative study of metagenomics and total RNA sequencing. Methods in Ecology and Evolution, 14(8), 2049–2064. https://doi.org/10.1111/2041-210X.14149
Raffington, J., Steinke, D., and Tulpan, D. (2020). Recognition of arthropod species names using bigram-based classification. In review. https://doi.org/10.21203/rs.3.rs-26532/v1
Sabanci, K., Kayabasi, A., and Toktas, A. (2017). Computer vision‐based method for classification of wheat grains using artificial neural network. Journal of the Science of Food and Agriculture, 97(8), 2588–2593. https://doi.org/10.1002/jsfa.8080
Steinke, D., McKeown, J. T. A., Zyba, A., McLeod, J., Feng, C., and Hebert, P. D. N. (2024). Low-cost, high-volume imaging for entomological digitization. ZooKeys. https://doi.org/10.3897/arphapreprints.e124163
Steinke, D., Ratnasingham, S., Agda, J. R. A., Boutou, H. A., Box, I. C. H., Boyle, M., Chan, D., Feng, C., Lowe, S. C., McKeown, J. T. A., McLeod, J., Sanchez, A., Smith, A., Walker, S., Wei, C. Y.-Y., and Hebert, P. D. N. (2024). Towards a taxonomy machine: A training set of 5.6 million arthropod images. Data 9 (11), 122. https://doi.org/10.3390/data9110122
Steinke, D., McKeown, J. T. A., Zyba, A., McLeod, J., Feng, C., and Hebert, P. D. N. (2024). Low-cost, high-volume imaging for entomological digitization. ZooKeys. https://doi.org/10.3897/arphapreprints.e124163
Steinke, D., Ratnasingham, S., Agda, J. R. A., Boutou, H. A., Box, I. C. H., Boyle, M., Chan, D., Feng, C., Lowe, S. C., McKeown, J. T. A., McLeod, J., Sanchez, A., Smith, A., Walker, S., Wei, C. Y.-Y., and Hebert, P. D. N. (2024). Towards a taxonomy machine: A training set of 5.6 million arthropod images. Data 9 (11), 122. https://doi.org/10.3390/data9110122
Dr. Dan Tulpan, Jaclyn McKeown, Dr. Joschka McLeod, Dr. Anton Potapov, Dr. Jerome Cortet, Dr. Jérôme Mathieu, Dr. Andrew MacDougall, Dr. Paul Fieguth, Dr. Angel Chang, Dr. Graham Taylor
Jennien Raffington, Julia Harvie, Aleksandra Dolezal, Nicholas Pellegrino, Christopher Hempel, Dr. Bettina Thalinger, Dr. Zahra Gharaee
University of Waterloo, Sorbonne University, University of Montpellier, Senckenberg Museum für Naturkunde GörlitzUniversity of Göttingen
