Department of Pathology, Stanford University School of Medicine, Stanford, CA, USA; Department of Biology, University of Copenhagen, Copenhagen, Denmark.
Department of Pathology, Stanford University School of Medicine, Stanford, CA, USA.
Cell Syst. 2024 Nov 20;15(11):1046-1060.e3. doi: 10.1016/j.cels.2024.10.004. Epub 2024 Nov 13.
Secreted proteins play crucial roles in paracrine and endocrine signaling; however, identifying ligand-receptor interactions remains challenging. Here, we benchmarked AlphaFold2 (AF2) as a screening approach to identify extracellular ligands to single-pass transmembrane receptors. Key to the approach is the optimization of AF2 input and output for screening ligands against receptors to predict the most probable ligand-receptor interactions. The predictions were performed on ligand-receptor pairs not used for AF2 training. We demonstrate high discriminatory power and a success rate of close to 90% for known ligand-receptor pairs and 50% for a diverse set of experimentally validated interactions. Further, we show that screen accuracy does not correlate linearly with prediction of ligand-receptor interaction. These results demonstrate a proof of concept of a rapid and accurate screening platform to predict high-confidence cell-surface receptors for a diverse set of ligands by structural binding prediction, with potentially wide applicability for the understanding of cell-cell communication.
分泌蛋白在旁分泌和内分泌信号中发挥着至关重要的作用；然而，鉴定配体-受体相互作用仍然具有挑战性。在这里，我们将 AlphaFold2（AF2）作为一种筛选方法进行基准测试，以识别针对单次跨膜受体的细胞外配体。该方法的关键是优化 AF2 的输入和输出，以筛选针对受体的配体，从而预测最可能的配体-受体相互作用。预测是针对未用于 AF2 训练的配体-受体对进行的。我们证明了针对已知配体-受体对的高区分能力和接近 90%的成功率，以及针对多样化的实验验证相互作用的 50%成功率。此外，我们表明，筛选准确性与配体-受体相互作用的预测并不呈线性相关。这些结果证明了一种快速而准确的筛选平台的概念验证，该平台通过结构结合预测，针对多样化的配体预测高可信度的细胞表面受体，具有广泛的应用潜力，可用于理解细胞间通讯。
