Article Details

Title · Authors · Abstract · Keywords · DOI · References

Research on Anti-breast Cancer Drug Activity Screening Based on Siamese Network

LI Yangfei1,2, WANG Qiumeng1,3, DAI Huarong4

(1.School of Artificial Intelligence, Mianyang Polytechnic, Mianyang 621000, China; 2.College of Life Science, China West Normal University, Nanchong 637009, China; 3.Mianyang Engineering Technological Research Center of Visual Object Detection and Recognition, Mianyang 621000, China; 4.Department of Thyroid Surgery, Mianyang Third People's Hospital, Mianyang 621000, China)

Section：Information Application · Vol. 2026, Issue 4 · Pages：138-145,152 · 2096-4706（2026）04-0138-09 · CLC number：TP18 · Document code：A · DOI：10.19850/j.cnki.2096-4706.2026.04.024

Abstract

Breast cancer is one of the most common malignant tumors in women, and precise and efficient screening methods for candidate drugs are urgently needed. This paper proposes a three-layer screening framework based on Siamese networks. Firstly, it filters key molecular descriptors through K-Means clustering and Mutual Information analysis to alleviate high-dimensional redundancy. Secondly, dual-branch multi-layer perceptrons are adopted to predict IC50_nM and pIC50 respectively, and the Siamese network is utilized to enhance small-sample ADMET data to improve classification performance. Finally, it realizes multi-objective optimization by combining Genetic Algorithms and Decision Tree rules. Experimental results show that the model reduces the activity prediction error by 15.2%, and the ADMET classification accuracy increases to 87.5%. The framework successfully screens 632 potential candidate molecules with both activity and drug-likeness, providing strong computational support for the early research and development of anti-breast cancer drugs.

Keywords

breast cancer; Estrogen  Receptors α (ERα); siamese network; bioactivity prediction; ADMET property optimization; candidate drug screening

References

  1. [1] CARDOSO F，O'SHAUGHNESSY J，LIU Z，et al. Pembrolizumab and Chemotherapy in High-risk, Early-stage, ER+/HER2− Breast Cancer: A Randomized Phase 3 Trial [J].Nature Medicine，2025：1-7.
  2. [2] XU Y C，HUANGYANG P W，WANG Y，et al. ERα is an RNA-Binding Protein Sustaining Tumor Cell Survival and Drug Resistance [J].Cell，2021，184（20）：5215-5229.
  3. [3] LIANG J，YAO X S，AOUAD P，et al. ERα Dysfunction Caused by ESR1 Mutations and Therapeutic Pressure Promotes Lineage Plasticity in ER+ Breast Cancer [J].Nature Cancer，2025：357-371.
  4. [4] SHERIDAN H M. Elacestrant: First Approval [J].Drugs，2023，83（6）：555-561.
  5. [5] ELLAWATTE G P，GANDHI H A，SESHADRI A，et al. A Perspective on Explanations of Molecular Prediction Models [J].Journal of Chemical Theory and Computation，2023，19（8）：2149-2160.
  6. [6] AN T，CHEN Y，CHEN Y，et al. A Machine Learning-based Approach to ERα Bioactivity and Drug ADMET Prediction [J].Frontiers in Genetics，2023：131087273-1087273.
  7. [7] EYNDRICKX W，LEWIS M，MORAWIETZ T，et al. MELLODDY: Cross-pharma Federated Learning at Unprecedented Scale Unlocks Benefits in QSAR Without Compromising Proprietary Information [J].Journal of Chemical Information and Modeling，2023：2331-2344.
  8. [8] DULSAT J，LóPEZ N B，ESTRADA T R，et al. Evaluation of Free Online ADMET Tools for Academic or Small Biotech Environments [J].Molecules，2023，28（2）：776-776.
  9. [9] Cha Jinmiao, Zheng Ziting, Yan Saihong Application, Development, and Prospects of Quantitative Structure Activity Relationship Models in the Field of Compound Toxicity Prediction [J]. Journal of Environmental Science, 2022, 42 (10): 1-11
  10. [10] JIA J，LEE G，WANG Z B，et al. Siamese Meets Diffusion Network: SMDNet for Enhanced Change Detection in High-Resolution RS Imagery [J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing，2024，17：8189-8202.
  11. [11] ONG W，HU Y，ZHANG J，et al. Application of SMILES-Based Molecular Generative Model in New Drug Design [J].Frontiers in Pharmacology，2022，13：1046524-1046524.
  12. [12] ANG Z X，PENG C Y，ZHANG Y，et al. Fully Convolutional Siamese networks Based Change Detection for Optical Aerial Images with focal Contrastive loss [J].Neurocomputing，2021，457：155-167.
  13. [13] BLOCKEEL H，DEVOS L，FRÉNAY B，et al. Decision Trees: from Efficient Prediction to Responsible AI [EB/OL].[2025-07-20].Frontiers in Artificial Intelligence，2023（6）：1124553[2025-09-28].https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2023.1124553/full.
  14. [14] Chinese Graduate Innovation Practice Series Competition The 18th Huawei Cup China Graduate Mathematical Modeling Competition [EB/OL]. [July 19, 2025] https://cpipc.aorg.cn//cw/detail/4/2c9080147c73b890017c7779e57e07d2.
  15. [15] HOU G F，RUSNAC D V，PARK H B，et al. An artificial Intelligence Accelerated Virtual Screening Platform for drug Discovery [J].Nature Communications，2024，15（1）：7761-7761.
  16. [16] LDAKHEEL F M，ALDURAYWISH S A，DABWAN K H. Integrating Machine Learning Driven Virtual Screening and Molecular Dynamics Simulations to Identify Potential Inhibitors Targeting PARP1 Against Prostate Cancer [J].Scientific Reports，2025，15（1）：12764-12764.
  17. [17] UAN N，MA S C，ZHAO K，et al. MFCADTI: Improving Drug-target Interaction Prediction by Integrating Multiple Feature Through Cross Attention Mechanism [J].BMC Bioinformatics，2025，26（1）：57.
  18. [18] ERREIRA L L，ANDRICOPULO A D. ADMET Modeling Approaches in Drug Discovery [J].Drug Discovery Today，2019，24（5）：1157-1165.

About the authors

Li Yangfei (2004-), male, Han ethnicity, from Mianyang, Sichuan, is currently pursuing a bachelor's degree with a research focus on biotechnology; Wang Qiumeng (1979-), male, Qiang ethnic group, from Mianyang, Sichuan, lecturer, master's degree, research direction: traffic information engineering and control (software); Dai Huarong (1979-), female, Han ethnicity, from Mianyang, Sichuan, is a nursing supervisor with a bachelor's degree and a research focus on nursing.