2023 SICBC-AI Algorithms for Drug Screening
举办方:STCSM,SCBD
举办方:STCSM,SCBD
Computational biology is an emerging interdisciplinary field that explores the structures, functions, and evolution of biological systems with an integrated approach derived from computer science, mathematics, statistics, and physics, among other fields. It has enormous potential for applications and high market value. In recent years, governments have given high priority to the development of computational biology, trying to solve biomedical challenges with computational biology-based methods and technologies.
As one of the regions with the greatest science and technology innovation strengths and vitality in biomedicine, Shanghai has a solid foundation for computational biology. However, the city still faces certain challenges, including the needs to allocate core resources, attract talents from all over the world, secure a leading position, ensure the high-quality development of computational biology, and further energize biomedicine in science and technology innovation and industrial development. To drive progress in computational biology, and implement Shanghai Action Plan for the Innovation-driven Development of Computational Biology (2023–2025), Shanghai International Computational Biology Challenge (SICBC), which focuses on algorithm development, model building, and AI-enabled drug design, is scheduled to be held in a row. The Competition aims to actively engage universities, scientific institutes, enterprises, and individuals from home and abroad, to select talents and projects, develop talent and project pools, and add Shanghai’s weight in domestic and international biological communities.
2023 Shanghai International Computational Biology Challenge —“Lingyue” Challenge of AI Algorithms for Drug Screening focuses on virtual drug screening targeting NMDA (N-methyl-D-aspartate) receptors . The Competition conducts real experiments to verify the virtual screening of drugs, providing an opportunity to objectively evaluate the practical capabilities of drug screening AI algorithms. The long-term goal of the Competition is to build an AI drug discovery platform that is highly recognized, credible, and influential in the industry, and to promote the discovery of original and innovative targets and drugs.
1. Raise innovation capabilities in computational biology to accelerate industrial development. Build a platform for resource matchmaking and ecological services to improve Shanghai’s innovation capabilities in computational biology, and exchange relevant resources including talents, technologies, funds, space, and third-party services for the accelerated the development of computational biology.
2. Select and attract outstanding talents to improve competitiveness. Attract talents in computational biology from all over the world, strengthen intersection of disciplines, engage in discussions on innovation frontiers, explore scenario-specific needs, and ignite sparks of innovation to improve Shanghai’s comprehensive strength and competitiveness in computational biology.
3. Develop the Competition into an internationally recognized brand to increase Shanghai’s presence in computational biology. Develop “Shanghai International Computational Biology Innovation Competition” into a brand to increase Shanghai’s international presence in computational biology.
4. Create a commercialization ecosystem of innovation applications to establish a network for resource matchmaking. Focus on innovative cutting-edge researches, application scenarios, and corresponding solutions, create a competition-based project and talent selection mechanism to attract more innovation, industrial, and capital investment resources to establish a resource matchmaking network, and facilitate the implementation of original technologies and products in various areas such as algorithm, software, and structure.
Clarification:
1. Individuals involved in question planning and event organization are ineligible for the Competition.
2. The competition account is for personal use by the participant only and shall not be transferred, rented, lent, sold, disclosed, or exposed to any third party in any form. The participant is responsible for ensuring the validity and authenticity of their account information. Consequences arising from invalid or wrong information shall be solely borne by the participant. The participant shall be solely accountable for the usage of the competition account. In the event of loss, theft, unauthorized use, or any other security issues concerning the account, it shall be reported to the organizing committee immediately.
1. Brief description of competition theme:
Identification of protein-ligand interactions is the core issue in drug design. In simple terms, a protein acts as a "lock" and a drug serves as the "key", with the drug binding with the protein based on the complementary pairing of shape and properties. Drug molecules that better suit the keyhole with high complementarity have increased activity and efficacy. Virtual screening aims to help design algorithms consistent with the keyhole structure (protein structure), enabling the identification of potential binders from a vast pool of candidate keys (molecules) for experimental validation, thus improving the efficiency and success rate of drug discovery.
The Competition focuses on NMDA (N-methyl-D-aspartate) receptors as the protein target. NMDA receptors are the promising therapeutic targets for various psychiatric and neurological disorders, such as stroke, depression, epilepsy, Alzheimer's disease, pain, etc. Developing drugs targeting NMDA receptors is of great clinical importance and market value. GluN1/GluN3A are subtypes of NMDA receptors which have not yet been widely developed. However, the limited availability of specific small molecule regulators and crystal structure information has greatly hindered the extensive exploration of GluN1/GluN3A receptors.
In this competition, participants are expected to use information on the protein sequences, structures, and active molecules of the NMDA receptor family to identify the potential highly-active molecules demonstrating specificity on GluN1/GluN3A receptors with an AI-based calculation method. Meanwhile, the Competition will combine will combine wet-lab experiments for verification to lay a crucial foundation for advanced drug development, and enable the emergence of valuable and reliable AI algorithms to facilitate AI-based drug development. (Navigate the “Competition Task” page for detailed description of competition theme)
2. Molecular compound libraries:
①The Compound Library, designated for this competition theme, contains about 18 million compounds and their detailed information, and the compounds obtained from the screening are purchased by the organizers;
②To foster the discovery of original new drugs, the Competition welcomes the submission of de novo generated molecule entities by the participating teams.
The Competition consists of 3 phases including registration, preliminary round and final:
(I) Registration (18:00, Dec. 14, 2023 to Feb. 29, 2024 [UTC+8])
Participants shall enter the Competition as individuals or teams (Team formation is mandatory for all participants), and submit entries within the specified time frame; the host will announce the advanced teams and molecules in March. Participating teams shall submit the following entries:
Item | Time | Requirements |
Submitting a project proposal | 18:00, Dec. 14, 2023 to Feb. 29, 2024 (UTC+8) | Submit a one-page project proposal. The host will allow no more than 100 participating teams to advance to the preliminary round considering research basis, comprehension of the competition theme, program feasibility and creativity, and other factors.(Refer to Attachment 4 at the bottom of the “Competition Task” page for the project proposal template) |
Submitting participating molecules | Submit a total of 300 participating molecules listed by predicted score in descending order. Each library shall contain 100 participating molecules for the preliminary round (Refer to Attachment 5 at the bottom of the “Competition Task” page for the submission template). The host will conduct wet-lab experiments to verify at least 10 molecules from each participating team in accordance with the priority of the moleculars submitted. | |
Submitting source code of participating molecules | The host encourages (but does not mandate) submission of the source code that can reproduce the participating molecules. |
(II) Preliminary Round (Mar. 01, 2024 to Aug. 31, 2024 [UTC+8])
In the preliminary round, the host will verify and evaluate the participating molecules through wet-lab experiment. Teams advanced to the final will be announced on August 31, 2024. The preliminary round consists of 3 phases:
Phase | Time | Schedule |
Preliminary round | Mar. 01, 2024 to Aug. 31, 2024 (UTC+8) | Shortlisting: The top 100 active molecules selected upon the results of high-throughput screening based on fluorescence signal will advance to the next phase. |
Preliminary assessment: The top 10 participating teams selected upon the results of preliminary comparison and activity verification based on current signal will advance to the next phase. | ||
Review: •Phase I: The top 5 participating teams selected upon the results of multi-concentration activity evaluation will advance to the next phase (and the final); five teams advanced to Phase I review but failed Phase II review will be awarded the Excellence Award. •Phase II: The host will announce the wet-lab experiment scores of the 5 participating teams advanced to the final upon the results of multi-target selective evaluation. |
(III) Final (09/2024)
The 5 participating teams advanced to the final will be invited to deliver an on-site oral defense, which will be scored by the jury of the final. The participating teams’ defense scores will be announced on site.
The host will tally weighted scores based on the participating teams’ wet-lab experiment scores (70%) and defense scores (30%) to determine the winners of the first, second and third prizes.
Note: Defense-related requirements and scoring criteria will be announced on the competition platform before the final. Participants shall stay tuned for official information.
1. Official communication forum: During the Competition, participating teams may post threads in the ModelArts Community (ModelArts Development and Discussion Forum). The thread shall be titled [2023 SICBC + Your Query], with a detailed description of your query in the body (screenshots may be attached). The organizing committee will reply to your thread on weekdays (9:00-17:00, UTC+8) through the community. Click here to view all your queries.
Clarification: For equality and fairness, the competition officials will only respond to answer queries about the usage of ModelArts, competition theme, data, scoring criteria, and schedule. Queries concerning code debugging won't be handled. For late replies, participating teams may post the thread’s URL in the discussion group for feedback.
2. Q&A group for participating teams: Please make sure to scan the chatbot’s QR code below, and send “2023SICBC” to join the discussion group for competition-related communication and Q&A. The organizing committee will promptly release important event information such as significant important milestones, live broadcasts, and award lists in the group.
Supervisor: Science and Technology Commission of Shanghai Municipality
Host: Shanghai Center of Biomedicine Development
Co-hosts: State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences; Bank of China Shanghai Branch; Huawei Cloud Computing Technologies Co., Ltd.; Lin Gang Laboratory; Shanghai Artificial Intelligence Research Institute;The Oriental Beauty Valley Enterprise Group Co., Ltd.; Shanghai Daning Asset Management (Group) Co., Ltd.
Organizers: Shanghai Biopharma Service Co., Ltd.; Shanghai Technological and Functional Platform for Biomedical Industry
Co-organizer: MedChemExpress (MCE)
The following cases are considered as violations of competition rules. The organizing committee shall have the right to disqualify violating teams from the Competition.
1. Providing false registration information or failing to comply with the Competition's registration or teaming requirements.
2. Suspected of plagiarism, intellectual property infringement, etc.
3. Other illegal behaviors or violations discovered or reported during the Competition or relevant to the entries submitted.
4. Competition data: The organizing committee only authorize participants to use the competition data for model training. Participants are strictly prohibited from employing the competition-based model training data for any commercial purpose.
5. Refer to “Attachment 1: Intellectual Property Agreement” at the bottom of the page for the competition conventions relevant to intellectual property.
The final interpretation of competition regulations shall be at the organizing committee’s discretion.