•We review machine learning methods/tools relevant to ligand-based virtual screening.•Machine learning methods classify compounds and predict new active molecules.•We discuss challenges, limitations ...and advantages of the methods and tools.•The wide applicability of the approaches is demonstrated in several case studies.•Some new algorithms and concepts in the machine learning field are provided.
During the past decade, virtual screening (VS) has evolved from traditional similarity searching, which utilizes single reference compounds, into an advanced application domain for data mining and machine-learning approaches, which require large and representative training-set compounds to learn robust decision rules. The explosive growth in the amount of public domain-available chemical and biological data has generated huge effort to design, analyze, and apply novel learning methodologies. Here, I focus on machine-learning techniques within the context of ligand-based VS (LBVS). In addition, I analyze several relevant VS studies from recent publications, providing a detailed view of the current state-of-the-art in this field and highlighting not only the problematic issues, but also the successes and opportunities for further advances.
This paper focuses on machine-learning approaches in the context of ligand-based virtual screening for addressing complex compound classification problems and predicting new active molecules.
•Deep learning methods have gained outstanding achievements.•We review deep learning methods/tools relevant to drug discovery research.•We discuss opportunities, challenges and advantages of methods ...and tools.•The wide applicability of the approaches is demonstrated in several case studies.•Future prospects of deep learning in drug discovery are discussed.
Artificial Intelligence (AI) is an area of computer science that simulates the structures and operating principles of the human brain. Machine learning (ML) belongs to the area of AI and endeavors to develop models from exposure to training data. Deep Learning (DL) is another subset of AI, where models represent geometric transformations over many different layers. This technology has shown tremendous potential in areas such as computer vision, speech recognition and natural language processing. More recently, DL has also been successfully applied in drug discovery. Here, I analyze several relevant DL applications and case studies, providing a detailed view of the current state-of-the-art in drug discovery and highlighting not only the problematic issues, but also the successes and opportunities for further advances.
This paper focuses on deep learning approaches in the context of drug discovery for designing new effective molecules, predicting for the desired molecular property profiles and planning synthesis.
Malignant melanoma is the most aggressive and life-threatening skin cancer. Melanoma develops in melanocytes and is characterized by a very high tendency to spread to other parts of the body. Its ...pathogenesis depends on DNA mutations leading to the activation of oncogenes or to the inactivation of suppressor genes. The identification of misregulations in intracellular signal transduction pathways has provided an opportunity for the development of mutation-specific inhibitors, which specifically target the mutated signaling cascades. Over the last few years, clinical trials with MAPK pathway inhibitors have shown significant clinical activity in melanoma; however, their efficacy is limited due to the onset of acquired resistance. This has prompted a large set of preclinical studies looking at new approaches of pathway- or target-specific inhibitors. This review gives an overview of the latest developments of small molecule targeting multiple molecular pathways in both preclinical and clinical melanoma settings, with particular emphasis on additional strategies to tackle the reduced responsiveness to inhibitor treatment as possible future directions.
•Revolutionizing drug discovery with deep attention neural networks.•Exploring attention mechanism and extended architectures like GATs, transformers, BERT, GPTs and BART for complex data.•Uncovering ...a pivotal role in catalyzing de novo drug design, predicting molecular properties and deciphering elusive drug–target interactions.•Addressing challenges to deepen understanding for pharmaceutical breakthroughs.
In the dynamic field of drug discovery, deep attention neural networks are revolutionizing our approach to complex data. This review explores the attention mechanism and its extended architectures, including graph attention networks (GATs), transformers, bidirectional encoder representations from transformers (BERT), generative pre-trained transformers (GPTs) and bidirectional and auto-regressive transformers (BART). Delving into their core principles and multifaceted applications, we uncover their pivotal roles in catalyzing de novo drug design, predicting intricate molecular properties and deciphering elusive drug–target interactions. Despite challenges, these attention-based architectures hold unparalleled promise to drive transformative breakthroughs and accelerate progress in pharmaceutical research.
•Explore how generative AI models navigate chemical space beyond structural constraints.•Examine RNNs, VAEs, GANs, NF models, and transformers for chemical space exploration.•Discuss challenges in ...molecular representations, training methods, and chemical space coverage criteria.•Future focus: refine models, adopt new notations, improve benchmarks, and enhance interpretability.
Deep generative models (GMs) have transformed the exploration of drug-like chemical space (CS) by generating novel molecules through complex, nontransparent processes, bypassing direct structural similarity. This review examines five key architectures for CS exploration: recurrent neural networks (RNNs), variational autoencoders (VAEs), generative adversarial networks (GANs), normalizing flows (NF), and Transformers. It discusses molecular representation choices, training strategies for focused CS exploration, evaluation criteria for CS coverage, and related challenges. Future directions include refining models, exploring new notations, improving benchmarks, and enhancing interpretability to better understand biologically relevant molecular properties.
Across life sciences, the steadily and rapidly increasing amount of data provide new opportunities for advancing knowledge and represent a key driver of emerging technological advancements ....
Systemic Sclerosis (SSc) is a heterogeneous autoimmune disease characterized by widespread vasculopathy, the presence of autoantibodies and the progressive fibrosis of skin and visceral organs. There ...are still many questions about its pathogenesis, particularly related to the complex regulation of the fibrotic process, and to the factors that trigger its onset. Our recent studies supported a key role of N-formyl peptide receptors (FPRs) and their crosstalk with uPAR in the fibrotic phase of the disease. Here, we found that dermal fibroblasts acquire a proliferative phenotype after the activation of FPRs and their interaction with uPAR, leading to both Rac1 and ERK activation, c-Myc phosphorylation and Cyclin D1 upregulation which drive cell cycle progression. The comparison between normal and SSc fibroblasts reveals that SSc fibroblasts exhibit a higher proliferative rate than healthy control, suggesting that an altered fibroblast proliferation could contribute to the initiation and progression of the fibrotic process. Finally, a synthetic compound targeting the FPRs/uPAR interaction significantly inhibits SSc fibroblast proliferation, paving the way for the development of new targeted therapies in fibrotic diseases.
The peroxisome proliferator-activated receptors (PPARs) are nuclear receptors involved in the regulation of the metabolic homeostasis and therefore represent valuable therapeutic targets for the ...treatment of metabolic diseases. The development of more balanced drugs interacting with PPARs, devoid of the side-effects showed by the currently marketed PPARγ full agonists, is considered the major challenge for the pharmaceutical companies. Here we present a structure-based virtual screening approach that let us identify a novel PPAR pan-agonist with a very attractive activity profile and its crystal structure in the complex with PPARα and PPARγ, respectively. In PPARα this ligand occupies a new pocket whose filling is allowed by the ligand-induced switching of the F273 side chain from a closed to an open conformation. The comparison between this pocket and the corresponding cavity in PPARγ provides a rationale for the different activation of the ligand towards PPARα and PPARγ, suggesting a novel basis for ligand design.
The cell division cycle 25 (CDC25) phosphatases include CDC25A, CDC25B and CDC25C. These three molecules are important regulators of several steps in the cell cycle, including the activation of ...various cyclin-dependent kinases (CDKs). CDC25s seem to have a role in the development of several human malignancies, including acute myeloid leukemia (AML); and CDC25 inhibition is therefore considered as a possible anticancer strategy. Firstly, upregulation of CDC25A can enhance cell proliferation and the expression seems to be controlled through PI3K-Akt-mTOR signaling, a pathway possibly mediating chemoresistance in human AML. Loss of CDC25A is also important for the cell cycle arrest caused by differentiation induction of malignant hematopoietic cells. Secondly, high CDC25B expression is associated with resistance against the antiproliferative effect of PI3K-Akt-mTOR inhibitors in primary human AML cells, and inhibition of this isoform seems to reduce AML cell line proliferation through effects on NFκB and p300. Finally, CDC25C seems important for the phenotype of AML cells at least for a subset of patients. Many of the identified CDC25 inhibitors show cross-reactivity among the three CDC25 isoforms. Thus, by using such cross-reactive inhibitors it may become possible to inhibit several molecular events in the regulation of cell cycle progression and even cytoplasmic signaling, including activation of several CDKs, through the use of a single drug. Such combined strategies will probably be an advantage in human cancer treatment.
•AI has gained much attention because of the recent achievements in biomedical research.•We review AI methods and tools relevant to drug discovery.•We highlight their wide applicability in several ...case studies.•We discuss potential future directions and opportunities for improvements.
Over the past decade, the amount of biomedical data available has grown at unprecedented rates. Increased automation technology and larger data volumes have encouraged the use of machine learning (ML) or artificial intelligence (AI) techniques for mining such data and extracting useful patterns. Because the identification of chemical entities with desired biological activity is a crucial task in drug discovery, AI technologies have the potential to accelerate this process and support decision making. In addition, the advent of deep learning (DL) has shown great promise in addressing diverse problems in drug discovery, such as de novo molecular design. Herein, we will appraise the current state-of-the-art in AI-assisted drug discovery, discussing the recent applications covering generative models for chemical structure generation, scoring functions to improve binding affinity and pose prediction, and molecular dynamics to assist in the parametrization, featurization and generalization tasks. Finally, we will discuss current hurdles and the strategies to overcome them, as well as potential future directions.