Using Hybrid Similarity-Based Collaborative Filtering Method for Compound Activity Prediction
It is important for researchers to predict compound activity to the targets quickly and effectively in the field of drug design. In the paper, the problem of compound activity prediction is converted to the recommendations in the field of e-commerce, compounds are viewed as users, and protein targets are viewed as items. A rating matrix is extracted by IC50 of each compound to targets, there are four filtering recommendation algorithms could be used for predicting compound activity. In order to improve the accuracy of prediction, the hybrid similarity-based Collaborative Filtering (HybridSimCF) Method is proposed, the method will combine the similarity of the compound structure and the similarity based on the rating matrix to predict the activity. Through compared with other three collaborative filtering methods, HybridSimCF has better results. It not only improves the values of RMSE and MAE, but also effectively solves the cold start problem. The method can quickly and effectively solve the prediction of compound activity.
KeywordsDrug design QSAR Compound activity Rating prediction Machine learning Collaborative filtering
This work was supported by the Fundamental Research Funds for the National Natural Science Foundation of China (Grant No. 21503101, No. 61702240), the Natural Science Foundation of Gansu Province, China (Grant No. 1506RJZA223), the Project Sponsored by the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry (Grant No. External department of Education  311) and the Fundamental Research Funds for the Central Universities (Grant No. lzujbky-2017-191).
- 5.Dahl, G.E., Jaitly, N., Salakhutdinov, R.: Multi-task Neural Networks for QSAR Predictions. Computer Science (2014)Google Scholar
- 7.Dearden, J.C., Rowe, P.H.: Use of artificial neural networks in the QSAR prediction of physicochemical properties and toxicities for REACH legislation. In: Cartwright, H. (ed.) Artificial Neural Networks. MMB, vol. 1260, pp. 65–88. Springer, New York (2015). https://doi.org/10.1007/978-1-4939-2239-0_5CrossRefGoogle Scholar
- 16.Tebby, C., Mombelli, E.: A kernel-based method for assessing uncertainty on individual QSAR predictions. QSAR Comb. Sci. 31, 741–751 (2015)Google Scholar
- 18.Erhan, D.: Collaborative filtering techniques for drug discovery (2006)Google Scholar
- 24.Xia, N.: Machine learning and data mining methods for recommender systems and chemical informatics. University of Minnesota (2012)Google Scholar
- 25.DRAGON Homepage: http://www.talete.mi.it/
- 26.Schafer, J.B., Konstan, J., Riedl, J.: Recommender systems in e-commerce. In: Proceedings of the 1st ACM Conference on Electronic Commerce, vol. 158 (1999)Google Scholar
- 27.Shi, J., Chen, J., Bao, Z.: An application study on collaborative filtering in e-commerce. In: International Conference on Service Systems and Service Management, vol. 1 (2011)Google Scholar
- 29.You, H., Li, H., Wang, Y., Zhao, Q.: An improved collaborative filtering recommendation algorithm combining item clustering and slope one scheme. Lect. Notes Eng. Comput. Sci. 2215, 313–316 (2015)Google Scholar
- 30.Sedhain S., Braziunas D., Braziunas D., Christensen J., Christensen J.: Social collaborative filtering for cold-start recommendations. In: ACM Conference on Recommender Systems, vol. 345 (2014)Google Scholar