Click the blue text to follow us
Product Overview
Protein-Protein Interactions (PPI) play a critical role in the biology of plants and animals, driving many cellular functions and biological processes. Therefore, identifying interactions between proteins is crucial for understanding biological mechanisms and developing drug and biotechnology applications. Traditional screening methods (such as Y2H, COIP, etc.) often take a long time, are inefficient, and require significant experimental resources. To improve screening efficiency, recent advancements in high-throughput virtual screening technologies and AI models have made structure- and sequence-based interaction predictions possible.
This product aims to utilize the cutting-edge generative model RFdiffusion to design artificial high-interaction proteins from scratch and combine advanced tools such as Foldseek, HDOCK, and AlphaFold 3 to screen and validate natural proteins. We named this new computational workflow InterProDesign (IPD, Interaction Protein Design). Through IPD, we can more efficiently and accurately screen natural proteins that interact with bait proteins, providing important references for further experimental validation.
Product Workflow
References indicate that this computational workflow has an experimental validation success rate exceeding78%, providing important structural biology information for studying the biological mechanisms of plants and animals.
1
RFdiffusion designs interaction proteins
Based on the known hotspot binding of bait proteins, it designs 10 high-interaction proteins of different lengths from scratch.
2
Foldseek searches for natural proteins
In the three-dimensional structure database of the target species, it compares and searches for structurally similar natural proteins (retaining only matches with bit score>50), and ranks them by score (up to the top 100).
3
Binding energy evaluation and overall structure assessment
For the high-scoring natural proteins selected in the previous round by Foldseek, the HDOCK tool is used to evaluate the quantitative binding energy between them and the bait protein, filtering for natural proteins with higher binding energy. Subsequently, AlphaFold 3 is used to further assess the confidence of the overall complex structure of these interaction proteins with the bait protein.
Scroll up to see more
Expected Results Illustration
NEW TECHNOLOGY
As shown in Figure 1, the data summary table includes the results of all steps. The red box contains the 10 interaction protein sequences designed by RFdiffusion based on the bait protein, and evaluates the docking of the artificial proteins with the bait protein.
Green box contains the structurally similar natural proteins obtained by Foldseek based on the designed proteins from the candidate species library. There are also natural proteins without a bit score greater than 50 for the artificial proteins.
Blue box contains the results of the quantitative binding energy and overall confidence assessment between the bait protein and the candidate natural proteins. The high-scoring proteins that stably interact have been highlighted in red, providing guidance for subsequent experimental validation.
As shown in Figure 2, detailed information of all candidate proteins is included, such as protein ID, gene name, sequence, species, subcellular localization, function annotation, and family.
Figure 1: Schematic diagram of the data summary table
Figure 2: Annotation information of candidate proteins
Product Advantages
NEW TECHNOLOGY
1
Innovative Computational Method
This product is based on a computational model for prediction and screening processes, which can significantly shorten screening time and reduce experimental workload. By combining tools such as RFdiffusion, Foldseek, HDOCK, and AlphaFold, it can quickly predict candidate proteins that may interact with the bait protein in the early stages. Multi-dimensional evaluations make the final selected interaction proteins more reliable, increasing the success rate of subsequent experimental validation.
2
Structure-Guided Precise Screening
This product relies more on the three-dimensional structure of proteins for screening. Structural similarity is crucial in protein interactions across species or with significant functional diversity. Structure-based screening methods can identify proteins that have low sequence similarity but exhibit similar functions.
References
1.Watson, J. L., et al. (2023). De novo design of protein structure and function with RFdiffusion. Nature, 620, 1089-1099.
2.Haley, O. C., Harding, S., Sen, T. Z., Woodhouse, M. R., Kim, H.-S., Andorf, C. (2024). Application of RFdiffusion to predict interspecies protein-protein interactions between fungal pathogens and cereal crops. bioRxiv, doi: 10.1101/2024.09.17.613523.
3.Watson, J. L., Juergens, D., Bennett, N. R., Trippe, B. L., Yim, J., Eisenach, H. E., … & Baker, D. (2023). De novo design of protein structure and function with RFdiffusion. Nature, 620(7685), 1089-1098.
4.Baek, M., DiMaio, F., Anishchenko, I., Dauparas, J., Ovchinnikov, S., Lee, G. R., … & Jumper, J. (2021). Accurate prediction of protein structures and interactions using a three-track neural network. Science, 373(6557), 871-876.
5.Lin, Z., Rigden, D. J., & McGuffin, L. J. (2023). Language models of protein sequences at the scale of evolution enable accurate structure prediction. Science, 379(6631), 1123-1130.
6.Dauparas, J., Anishchenko, I., Bennett, N. R., Bai, H., Ragotte, R. J., Milles, L. F., … & Baker, D. (2022). Robust deep learning-based protein sequence design using ProteinMPNN. Science, 378(6615), 49-56.
About KJ Biology
Hefei KJ Biology is committed to promoting biological development through computer virtual screening, helping to screen interaction proteins, transcription factors, potential drugs, and nanobodies.
Benefits are coming!
KJ Biology paper illustrations
Free protein interaction analysis awaits you!
0551-64669001
Scan the code to add customer service and receive!
▼
NEWS
