Current Medicinal Chemistry - Volume 28, Issue 32, 2021

Background: The adoption of biomarkers as part of high-throughput, complex microarray or sequencing data has necessitated the discovery and validation of these data through machine learning. Machine learning has remained a fundamental and indispensable tool due to its efficacy and efficiency in both feature extraction of relevant biomarkers as well as the classification of samples as validation of the discovered biomarkers. Objectives: This review aims to present the impact and ability of various machine learning methodologies and models to process high-throughput, high-dimensionality data found within mass spectrometry, microarray, and DNA/RNA-sequence data; data that precluded biomarker discovery prior to the use of machine learning. Methods: A vast array of literature highlighting machine learning for biomarker discovery was reviewed, resulting in the eligibility of 21 machine learning algorithms/networks and 3 combinatory architectures, spanning 17 fields of study. This literature was screened to investigate the usage and development of machine learning within the framework of biomarker discovery. Results: Out of the 93 papers collected, a total of 62 biomarker studies were further reviewed across different subfields-49 of which employed machine learning algorithms, and 13 of which employed neural network-based models. Through the application, innovation, and creation of tools in biomarker-related machine learning methodologies, its use allowed for the discovery, accumulation, validation, and interpretation of biomarkers within varied data formats, sources, as well as fields of study. Conclusion: The use of machine learning methodologies for biomarker discovery is critical to the analysis of various types of data used for biomarker discovery, such as mass spectrometry, nucleotide and protein sequencing, and image (e.g. CT-scan) data. Further studies containing more standardized techniques for evaluation, and the use of cutting- edge machine learning architectures may lead to more accurate and specific results.

Current methods for diagnosing human disease are still incapable of rapidly and accurately screening for multiple diseases simultaneously on a large scale, and at an affordable price. MALDI-ToF mass spectrometry is an ultra-sensitive, ultra-fast, lowcost, high-throughput technology that has the potential to achieve this goal, allowing human phenotype characterization and thus phenomic screening for multiple disease states. In this review, we will discuss the main advances achieved so far, putting forward targeted applications of MALDI-ToF mass spectrometry in the service of human disease detection. This review focuses on the methodological workflow as MALDI-ToF data processing for phenomic analysis, using state-of-the-art bioinformatic pipelines and software tools. The role of mathematical modelling, machine learning, and artificial intelligence algorithms for disease screening are considered. Moreover, we present some previously developed tools for disease diagnostics and screening based on MALDI-ToF analysis. We discuss the remaining challenges that are ahead when implementing MALDI-ToF into clinical laboratories. Differentiating a standard profile from a single disease phenotype is challenging, but the potential to simultaneously run multiple algorithm screens for different disease phenotypes may only be limited by computing power once this initial hurdle is overcome. The ability to explore the full potential of human clinical phenomics may be closer than imagined; this review gives an insight into the benefits this technology may reap for the future of clinical diagnostics.

Background: Parkinson’s disease is the second most frequent neurodegenerative disorder. Its diagnosis is challenging and mainly relies on clinical aspects. At present, no biomarker is available to obtain a diagnosis of certainty in vivo. Objective: The present review aims at describing machine learning algorithms as they have been variably applied to different aspects of Parkinson’s disease diagnosis and characterization. Methods: A systematic search was conducted on PubMed in December 2019, resulting in 230 publications obtained with the following search query: “Machine Learning” “AND” “Parkinson Disease”. Results: The obtained publications were divided into 6 categories, based on different application fields: “Gait Analysis - Motor Evaluation”, “Upper Limb Motor and Tremor Evaluation”, “Handwriting and typing evaluation”, “Speech and Phonation evaluation”, “Neuroimaging and Nuclear Medicine evaluation”, “Metabolomics application”, after excluding the papers of general topic. As a result, a total of 166 articles were analyzed after elimination of papers written in languages other than English or not directly related to the selected topics. Conclusion: Machine learning algorithms are computer-based statistical approaches that can be trained and are able to find common patterns from big amounts of data. The machine learning approaches can help clinicians in classifying patients according to several variables at the same time.

Background: In the real world, medical practice is changing hand in hand with the development of new Artificial Intelligence (AI) systems and problems from different areas have been successfully solved using AI algorithms. Specifically, the use of AI techniques in setting up or building precision medicine is significant in terms of the accuracy of disease discovery and tailored treatment. Moreover, with the use of technology, clinical personnel can deliver a very much efficient healthcare service. Objective: This article reviews AI state-of-the-art in cardiovascular disease management, focusing on diagnostic and therapeutic improvements. Methods: To that end, we conducted a detailed PubMed search on AI application from distinct areas of cardiology: heart failure, arterial hypertension, atrial fibrillation, syncope and cardiovascular rehabilitation. Particularly, to assess the impact of these technologies in clinical decision-making, this research considers technical and medical aspects. Results: On one hand, some devices in heart failure, atrial fibrillation and cardiac rehabilitation represent an inexpensive, not invasive or not very invasive approach to long-term surveillance and management in these areas. On the other hand, the availability of large datasets (big data) is a useful tool to predict the development and outcome of many cardiovascular diseases. In summary, with this new guided therapy, the physician can supply prompt, individualised, and tailored treatment and the patients feel safe as they are continuously monitored, with a significant psychological effect. Conclusion: Soon, tailored patient care via telemonitoring can improve clinical practice because AI-based systems support cardiologists in daily medical activities, improving disease detection and treatment. However, the physician-patient relationship remains a pivotal step.

Autistic Spectrum Disorder (ASD) is a neurodevelopmental condition affecting approximately 1 out of 70 (range 1:59 – 1:89) children worldwide. It is characterized by a delay in cognitive capabilities, repetitive and restricted behaviors and deficit in communication and social interaction. Several factors seem to be associated with ASD development; its heterogeneous nature makes the diagnosis difficult and slow since it is essentially based on screening tools focused on stereotypical and repetitive behaviors, gait, facial emotion expression and speech assessments. Recently, artificial intelligence (AI) has been widely used to investigate ASD with the overall goal of simplifying and speeding up the diagnostic process as well as making earlier access to therapies possible. The aim of this review is to provide an overview of the state-of-the-art research in the ASD field, identifying and describing machine learning (ML) approaches in ASD literature that could be used by clinicians to improve diagnostic capability and treatment efficiency. A systematic search was conducted and the resulting articles were subdivided into several categories reflecting the different fields of study associated with ASD research. The existing literature has widely demonstrated the potential of ML in several types of ASD study analyses: behavior, gait, speech, facial emotion expression, neuroimaging, genetics, and metabolomics. Therefore, AI techniques are becoming increasingly implemented and accepted, so highlighting the power of ML approaches to extract and obtain knowledge from a large volume of data. This makes ML a promising tool for future ASD research and clinical endeavors suggesting possible avenues for improving ASD screening, diagnostic and therapeutic tools.

Background: Systems biology and network modeling represent, nowadays, the hallmark approaches for the development of predictive and targeted-treatment based precision medicine. The study of health and disease as properties of the human body system allows the understanding of the genotype-phenotype relationship through the definition of molecular interactions and dependencies. In this scenario, metabolism plays a central role as its interactions are well characterized and it is considered an important indicator of the genotype- phenotype associations. In metabolic systems biology, the genome-scale metabolic models are the primary scaffolds to integrate multi-omics data as well as cell-, tissue-, condition- specific information. Modeling the metabolism has both investigative and predictive values. Several methods have been proposed to model systems, which involve steady-state or kinetic approaches, and to extract knowledge through machine and deep learning. Methods: This review collects, analyzes, and compares the suitable data and computational approaches for the exploration of metabolic networks as tools for the development of precision medicine. To this extent, we organized it into three main sections: "Data and Databases", "Methods and Tools", and "Metabolic Networks for medicine". In the first one, we have collected the most used data and relative databases to build and annotate metabolic models. In the second section, we have reported the state-of-the-art methods and relative tools to reconstruct, simulate, and interpret metabolic systems. Finally, we have reported the most recent and innovative studies that exploited metabolic networks to study several pathological conditions, not only those directly related to metabolism. Conclusion: We think that this review can be a guide to researchers of different disciplines, from computer science to biology and medicine, in exploring the power, challenges and future promises of the metabolism as predictor and target of the so-called P4 medicine (predictive, preventive, personalized and participatory).

Epigenetics is a field of biological sciences focused on the study of reversible, heritable changes in gene function, not due to modifications of the genomic sequence. These changes are the result of a complex cross-talk between several molecular mechanisms that is in turn orchestrated by genetic and environmental factors. The epigenetic profile captures the unique regulatory landscape and the exposure to environmental stimuli of an individual. It thus constitutes a valuable reservoir of information for personalized medicine, which is aimed at customizing health-care interventions based on the unique characteristics of each individual. Nowadays, the complex milieu of epigenomic marks can be studied at the genome-wide level thanks to massive, high-throughput technologies. This new experimental approach is opening up new and interesting knowledge perspectives. However, the analysis of these complex omic data requires to face important analytic issues. Artificial Intelligence, and in particular Machine Learning, are emerging as powerful resources to decipher epigenomic data. In this review, we will first describe the most used ML approaches in epigenomics. We then will recapitulate some of the recent applications of ML to epigenomic analysis. Finally, we will provide some examples of how the ML approach to epigenetic data can be useful for personalized medicine.

Background: Pharmacogenetics investigates the response to pharmacological treatments based on individual genetic background. Actually, numerous pharmacogenetic tests help to predict the response to drugs used in different medical areas, contributing to the so-called personalized medicine. Objective: This review aims to update the available data on the genotype-guided treatment with both the anticoagulant and antiplatelet agents. Moreover, it sheds light on the pitfalls in the implementation of cardiovascular pharmacogenetics. Methods: A review of the literature on the studies investigating the effects of the genotype- guided anticoagulant and antiplatelet treatment was performed. Results: Considering the extensive use of antithrombotic drugs, pharmacogenetics has particular importance in this field. Several polymorphisms influence the response to both anticoagulant and antiplatelet agents, and tests, based on their identification, are now available. Conclusion: Recent randomized clinical trials demonstrated that pharmacogenetics might successfully contribute to optimizing the antiplatelet therapy also in patients particularly complicated to treat. However, despite accumulating evidence on the utility and feasibility of some pharmacogenetics tests, several barriers still exist in their implementation in clinical practice.

Artificial intelligence methods, in particular, machine learning, has been playing a pivotal role in drug development, from structural design to the clinical trial. This approach is harnessing the impact of computer-aided drug discovery due to large available data sets for drug candidates and its new and complex manner of information interpretation to identify patterns for the study scope. In the present review, recent applications related to drug discovery and therapies are assessed, and limitations and future perspectives are analyzed.