Similarity-Based Virtual Screen Using Enhanced Siamese Deep Learning Methods

Abstract

Traditional drug production is a long and complex process that leads to new drug production. The virtual screening technique is a computational method that allows chemical compounds to be screened at an acceptable time and cost. Several databases contain information on various aspects of biologically active substances. Simple statistical tools are difficult to use because of the enormous amount of information and complex data samples of molecules that are structurally heterogeneous recorded in these databases. Many techniques for capturing the biological similarity between a test compound and a known target ligand in LBVS have been established. However, despite the good performances of the above methods compared to their prior, especially when dealing with molecules that have homogeneous active structural elements, they are not satisfied when dealing with molecules that are structurally heterogeneous. Deep learning models have recently achieved considerable success in a variety of disciplines due to their powerful generalization and feature extraction capabilities. Also, the Siamese network has been used in similarity models for more complicated data samples, especially with heterogeneous data samples. The main aim of this study is to enhance the performance of similarity searching, especially with molecules that are structurally heterogeneous. The Siamese architecture will be enhanced using two similarity distance layers with one fusion layer to further improve the similarity measurements between molecules and then adding many layers after the fusion layer for some models to improve the retrieval recall. In this architecture, several methods of deep learning have been used, which are long short-term memory (LSTM), gated recurrent unit (GRU), convolutional neural network-one dimension (CNN1D), and convolutional neural network-two dimensions (CNN2D). A series of experiments have been carried out on real-world data sets, and the results have shown that the proposed methods outperformed the existing methods.

Leveraging bounded datapoints to classify molecular potency improvements

“…As such, they use a distance function for locality-sensitive hashing as a contrastive learning approach. Siamese neural networks have been applied within the field of drug discovery to predict molecular similarity, 11 bioactivity, 12 toxicity, 13 drug–drug interactions, 14 relative free energy of binding, 15 and transcriptional response similarity. 16 These models have additionally shown particular promise when trained only on compounds with high similarity, highlighting how additional preprocessing steps, such as reducing exhaustive pairing to only the most similar pairs, can reduce computational costs while preserving predictive power for paired models.…”

Deep Learning Based Methods for Molecular Similarity Searching: A Systematic Review

“…The effect of different weighting schemes on the utility of molecular similarity measures has been the subject of interesting studies [7,127]. In addition to the further extended literature related to the structure descriptors [81,128] and the similarity coefficients [1,3,24], several research articles have employed weighting schemes to improve the recall and accuracy performances [7,8,77,127,129,130]. It is based on this concept that non-related molecular fragments weigh equally to the relevant fragments in terms of biological activity.…”

“…To enhance the effectiveness of molecular similarity searching, there is a need to consider the use of more reliable feature representations and develop deeper architectures [7,8]. Therefore, developing deep-learning-based models on how these important features are explored to improve the effectiveness of the similarity measure performance becomes a promising solution.…”

“…Considering the recent successes of deep-learning-based models in various application domains comprised of natural language processing (NLP) [4], image processing [5], and machine translations [6], deep learning models have been used for molecular similarity searching, which improves the capabilities by resolving several issues with the current models [7][8][9]. Particularly, the main idea of deep-learning-based approaches is based on how these models can be best utilized to select only the important features and how these important features can be used to improve the efficiency of molecular similarity searching.…”