Virtual screening by a new Clustering-based Weighted Similarity Extreme Learning Machine approach

Abstract

Machine learning techniques are becoming popular in virtual screening tasks. One of the powerful machine learning algorithms is Extreme Learning Machine (ELM) which has been applied to many applications and has recently been applied to virtual screening. We propose the Weighted Similarity ELM (WS-ELM) which is based on a single layer feed-forward neural network in a conjunction of 16 different similarity coefficients as activation function in the hidden layer. It is known that the performance of conventional ELM is not robust due to random weight selection in the hidden layer. Thus, we propose a Clustering-based WS-ELM (CWS-ELM) that deterministically assigns weights by utilising clustering algorithms i.e. k-means clustering and support vector clustering. The experiments were conducted on one of the most challenging datasets-Maximum Unbiased Validation Dataset-which contains 17 activity classes carefully selected from PubChem. The proposed algorithms were then compared with other machine learning techniques such as support vector machine, random forest, and similarity searching. The results show that CWS-ELM in conjunction with support vector clustering yields the best performance when utilised together with Sokal/Sneath(1) coefficient. Furthermore, ECFP_6 fingerprint presents the best results in our framework compared to the other types of fingerprints, namely ECFP_4, FCFP_4, and FCFP_6.

Fig 1. Architecture of the WS-ELM.

Fig 2. Relative improvement/worsening with respect to similarity searching for top 1% retrieved–average across ten runs, 16 similarity coefficients, and four fingerprints.

Fig 3. Violin plot of maximum percentage of active molecules retrieved in the top 1% with WS-ELM in conjunction with 16 different similarity coefficients–averaged across ten runs, 17 activity classes, and four fingerprints.

Fig 4. Maximum percentage of active molecules retrieved in the top 1% with WS-ELM and similarity searching in 17 activity classes–averaged across ten runs, 16 similarity coefficients, and four fingerprints.

Fig 5. Maximum percentage of active molecules retrieved with WS-ELM and similarity searching using four different fingerprints–averaged across 17 activity classes, 16 similarity coefficients, and 10 runs.

Fig 6. Effect of AUROC when the number of hidden nodes in WS-ELM and CWS-ELM_KMC is changed in activity class I01.

Solid lines represent mean values while shaded areas represent error/confidence bounds. The upper and lower bounds of each node are based on the standard deviation.

Fig 7. Effect of AUROC when the number of hidden nodes in WS-ELM and CWS-ELM_KMC is changed in activity class I17.

Solid lines represent mean values while shaded areas represent error/confidence bounds. The upper and lower bounds of each node are based on the standard deviation.

Fig 8. Enrichment plot for the top 1% of the sorted library for each performer with ECFP_6 fingerprint on activity class I01.

Fig 9. Enrichment plot for the top 1% of the sorted library for each performer with ECFP_6 fingerprint on activity class I17.

Fig 10. Molecules retrieved by different methods in top 1% of the ranked database for activity class I01.

Fig 11. Molecules retrieved by different methods in top 1% of the ranked database for activity class I17.

Fig 12. Early recognition criteria suggested by [35, 38].

(Left) EF (Right) Ratio of true positive rate to the false positive rate, at 0.5%, 1.0%, 2.0%, and 5.0% of the ranked database for WS-ELM and its variants, SVM, RF, and Similarity Searching (SS). Each bar represents the mean value across all activity classes and ten runs.

Fig 13. Bar charts showing mean EF and BEDROC at 1.0% of the ranked database for WS-ELM and its variants, SVM, RF, and Similarity Seaching (SS).

According to Truchon & Bayly, the top 1% of the ranked database is equivalent to α = 160.9 of BEDROC [34].