{"id":1493,"date":"2026-07-07T11:47:20","date_gmt":"2026-07-07T09:47:20","guid":{"rendered":"https:\/\/bdagroup.nl\/?post_type=internship&#038;p=1493"},"modified":"2026-07-07T11:52:07","modified_gmt":"2026-07-07T09:52:07","slug":"agentic-reasoning-for-terpene-synthase-product-prediction-guiding","status":"publish","type":"internship","link":"https:\/\/bdagroup.nl\/?internship=agentic-reasoning-for-terpene-synthase-product-prediction-guiding","title":{"rendered":"Agentic reasoning for Terpene synthase product prediction guiding"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">Genomic datasets hold hundreds of thousands of uncharacterised sequences, only a handful of which can be tested in the lab. Datasets that size are beyond what any researcher can work through by hand, so most sequences are never seriously considered. Agentic AI systems can autonomously traverse large, heterogeneous datasets, reasoning over sequence, structure, and prior knowledge to decide what is worth taking a closer look (<a href=\"https:\/\/doi.org\/10.1039\/d4dd00013g\" target=\"_blank\" rel=\"noreferrer noopener\">Ghafarollahi &amp; Buehler, 2024<\/a>; <a href=\"https:\/\/doi.org\/10.48550\/arXiv.2511.19423\" target=\"_blank\" rel=\"noreferrer noopener\">Jacob et al., 2025<\/a>).\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Terpene synthases (TPS) are a vast enzyme family, and an ideal target for this approach. They convert a few simple substrates into thousands of distinct terpene scaffolds, the basis of many fragrances, flavours, and drugs (<a href=\"https:\/\/doi.org\/10.1021\/acs.chemrev.7b00287\" target=\"_blank\" rel=\"noreferrer noopener\">Christianson, 2017<\/a>). Their product cannot be read off from sequence; minor active-site differences can completely change the product outcome (<a href=\"https:\/\/doi.org\/10.3389\/fpls.2019.01166\" target=\"_blank\" rel=\"noreferrer noopener\">Karunanithi &amp; Zerbe, 2019<\/a>), so most of the TPS sequences accumulating in genomes have unknown, and potentially novel, product chemistry.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In this project you will build agentic reasoning workflows for enzyme discovery, aimed at nominating TPS candidates likely to produce novel or unusual terpene scaffolds. The agent will be given a set of tools, sequence search, structure prediction and retrieval, domain and motif detection, function classifiers, and literature retrieval, and tasked with weighing this evidence to decide which sequences are worth pursuing and to justify each choice.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Concretely, you will (I) design and implement the agentic workflow, covering planning, tool use, candidate ranking, and written justification; (II) benchmark it against simpler baselines [e.g. sequence similarity search] to confirm it adds real value; and (III) hand off a shortlist of nominated candidates for experimental product characterisation.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><\/h2>\n","protected":false},"template":"","categories":[],"tags":[],"class_list":["post-1493","internship","type-internship","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/bdagroup.nl\/index.php?rest_route=\/wp\/v2\/internship\/1493","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/bdagroup.nl\/index.php?rest_route=\/wp\/v2\/internship"}],"about":[{"href":"https:\/\/bdagroup.nl\/index.php?rest_route=\/wp\/v2\/types\/internship"}],"wp:attachment":[{"href":"https:\/\/bdagroup.nl\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1493"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/bdagroup.nl\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1493"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/bdagroup.nl\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1493"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}