Genome-wide prediction and prioritization of human aging genes by data fusion: a machine learning approach

Developmental dynamics of the single nucleus regulatory landscape of pig hippocampus

The hippocampus is a brain region associated with memory, learning and spatial navigation, its aging-related dysfunction is a common sign of Alzheimer's disease. Pig is a good model for human neurodegenerative disease, but our understanding of the regulatory program of the pig hippocampus and its cross-species conservation in humans remains limited. Here, we profiled chromatin accessibility in 33,409 high-quality nuclei and gene expression in 8,122 high-quality nuclei of the pig hippocampus at four postnatal stages. We identified 510,908 accessible chromatin regions (ACRs) in 12 major cell types, among which progenitor cells such as neuroblasts and oligodendrocyte progenitor cells showed a dynamic decrease from early to later developmental stages. We revealed significant enrichment of transposable elements in cell type-specific ACRs, particularly in neuroblasts. We identified oligodendrocytes as the most prominent cell type with the greatest number of genes that showed significant changes during the development. We identified ACRs and key transcription factors underlying the trajectory of neurogenesis (such as POU3F3 and EGR1) and oligodendrocyte differentiation (RXRA and FOXO6). We examined 27 Alzheimer's disease-related genes in our data and found that 15 showed cell type-specific activity (TREM2, RIN3 and CLU), and 15 genes displayed age-associated dynamic activity (BIN1, RABEP1 and APOE). We intersected our data with human genome-wide association study results to detect neurological disease-associated cell types. The present study provides a single nucleus-accessible chromatin landscape of the pig hippocampus at different developmental stages and is helpful for the exploration of pigs as a biomedical model in human neurodegenerative diseases.

Detecting biased validation of predictive models in the positive-unlabeled setting: disease gene prioritization case study

Motivation

Positive-unlabeled data consists of points with either positive or unknown labels. It is widespread in medical, genetic, and biological settings, creating a high demand for predictive positive-unlabeled models. The performance of such models is usually estimated using validation sets, assumed to be selected completely at random (SCAR) from known positive examples. For certain metrics, this assumption enables unbiased performance estimation when treating positive-unlabeled data as positive/negative. However, the SCAR assumption is often adopted without proper justifications, simply for the sake of convenience.

Results

We provide an algorithm that under the weak assumptions of a lower bound on the number of positive examples can test for the violation of the SCAR assumption. Applying it to the problem of gene prioritization for complex genetic traits, we illustrate that the SCAR assumption is often violated there, causing the inflation of performance estimates, which we refer to as validation bias. We estimate the potential impact of validation bias on performance estimation. Our analysis reveals that validation bias is widespread in gene prioritization data and can significantly overestimate the performance of models. This finding elucidates the discrepancy between the reported good performance of models and their limited practical applications.

Availability and implementation

Python code with examples of application of the validation bias detection algorithm is available at github.com/ArtomovLab/ValidationBias.

Synergy between machine learning and natural products cheminformatics: Application to the lead discovery of anthraquinone derivatives

Cheminformatics utilizing machine learning (ML) techniques have opened up a new horizon in drug discovery. This is owing to vast chemical space expansion with rocketing numbers of expected hits and lead compounds that match druggable macromolecular targets, in particular from natural compounds. Due to the natural products' (NP) structural complexity, uniqueness, and diversity, they could occupy a bigger space in pharmaceuticals, allowing the industry to pursue more selective leads in the nanomolar range of binding affinity. ML is an essential part of each step of the drug design pipeline, such as target prediction, compound library preparation, and lead optimization. Notably, molecular mechanic and dynamic simulations, induced docking, and free energy perturbations are essential in predicting best binding poses, binding free energy values, and molecular mechanics force fields. Those applications have leveraged from artificial intelligence (AI), which decreases the computational costs required for such costly simulations. This review aimed to describe chemical space and compound libraries related to NPs. High-throughput screening utilized for fractionating NPs and high-throughput virtual screening and their strategies, and significance, are reviewed. Particular emphasis was given to AI approaches, ML tools, algorithms, and techniques, especially in drug discovery of macrocyclic compounds and approaches in computer-aided and ML-based drug discovery. Anthraquinone derivatives were discussed as a source of new lead compounds that can be developed using ML tools for diverse medicinal uses such as cancer, infectious diseases, and metabolic disorders. Furthermore, the power of principal component analysis in understanding relevant protein conformations, and molecular modeling of protein-ligand interaction were also presented. Apart from being a concise reference for cheminformatics, this review is a useful text to understand the application of ML-based algorithms to molecular dynamics simulation and in silico absorption, distribution, metabolism, excretion, and toxicity prediction.

A Genome-Wide Association Study and Machine-Learning Algorithm Analysis on the Prediction of Facial Phenotypes by Genotypes in Korean Women

Purpose

Changes in facial appearance are affected by various intrinsic and extrinsic factors, which vary from person to person. Therefore, each person needs to determine their skin condition accurately to care for their skin accordingly. Recently, genetic identification by skin-related phenotypes has become possible using genome-wide association studies (GWAS) and machine-learning algorithms. However, because most GWAS have focused on populations with American or European skin pigmentation, large-scale GWAS are needed for Asian populations. This study aimed to evaluate the correlation of facial phenotypes with candidate single-nucleotide polymorphisms (SNPs) to predict phenotype from genotype using machine learning.

Materials and Methods

A total of 749 Korean women aged 30–50 years were enrolled in this study and evaluated for five facial phenotypes (melanin, gloss, hydration, wrinkle, and elasticity). To find highly related SNPs with each phenotype, GWAS analysis was used. In addition, phenotype prediction was performed using three machine-learning algorithms (linear, ridge, and linear support vector regressions) using five-fold cross-validation.

Results

Using GWAS analysis, we found 46 novel highly associated SNPs (p < 1×10⁻⁰⁵): 3, 20, 12, 6, and 5 SNPs for melanin, gloss, hydration, wrinkle, and elasticity, respectively. On comparing the performance of each model based on phenotypes using five-fold cross-validation, the ridge regression model showed the highest accuracy (r² = 0.6422–0.7266) in all skin traits. Therefore, the optimal solution for personal skin diagnosis using GWAS was with the ridge regression model.

Conclusion

The proposed facial phenotype prediction model in this study provided the optimal solution for accurately predicting the skin condition of an individual by identifying genotype information of target characteristics and machine-learning methods. This model has potential utility for the development of customized cosmetics.

Artificial intelligence and machine learning approaches for drug design: challenges and opportunities for the pharmaceutical industries

The global spread of COVID-19 has raised the importance of pharmaceutical drug development as intractable and hot research. Developing new drug molecules to overcome any disease is a costly and lengthy process, but the process continues uninterrupted. The critical point to consider the drug design is to use the available data resources and to find new and novel leads. Once the drug target is identified, several interdisciplinary areas work together with artificial intelligence (AI) and machine learning (ML) methods to get enriched drugs. These AI and ML methods are applied in every step of the computer-aided drug design, and integrating these AI and ML methods results in a high success rate of hit compounds. In addition, this AI and ML integration with high-dimension data and its powerful capacity have taken a step forward. Clinical trials output prediction through the AI/ML integrated models could further decrease the clinical trials cost by also improving the success rate. Through this review, we discuss the backend of AI and ML methods in supporting the computer-aided drug design, along with its challenge and opportunity for the pharmaceutical industry. From the available information or data, the AI and ML based prediction for the high throughput virtual screening. After this integration of AI and ML, the success rate of hit identification has gained a momentum with huge success by providing novel drugs.

Identification of Repurpose Drugs by Computational Analysis of Disease-Gene-Drug Associations

Repurposing of marketed drugs to find new indications has become an alternative to circumvent the risk of traditional drug development by its productivity quality. Despite many approaches, computational analysis has great potential to fuel the development of all-rounder drugs to find new classes of medicine for neglected and rare disease. The genes that can explain variations in drug response associated to disease are more important and significant in drug therapeutics necessitate elucidating the relationships of a gene, drug, and disease. The proposed computational analysis facilitates the discovery of knowledge on both target and disease-based relationships from large sources of biomedical literature spread over different platforms. It uses the utility of text mining for automatic extraction of valuable aggregated biomedical entities (disease, gene, and drug) from PubMed to serves as an input to the analysis of association prediction. The top-ranked associations considered for identification of repurposing drugs and also the hidden associations identified using concurrence principle to extrapolate the new relationships. Such findings are reported as novel and contribute to the knowledge base for pharmacogenomics, would immensely support the discovery and progress of novel therapeutic pathways and patient segment biomarkers.

The Emerging Roles of the RNA Binding Protein QKI in Cardiovascular Development and Function

RNA binding proteins (RBPs) have a broad biological and physiological function and are critical in regulating pre-mRNA posttranscriptional processing, intracellular migration, and mRNA stability. QKI, also known as Quaking, is a member of the signal transduction and activation of RNA (STAR) family, which also belongs to the heterogeneous nuclear ribonucleoprotein K- (hnRNP K-) homology domain protein family. There are three major alternatively spliced isoforms, QKI-5, QKI-6, and QKI-7, differing in carboxy-terminal domains. They share a common RNA binding property, but each isoform can regulate pre-mRNA splicing, transportation or stability differently in a unique cell type-specific manner. Previously, QKI has been known for its important role in contributing to neurological disorders. A series of recent work has further demonstrated that QKI has important roles in much broader biological systems, such as cardiovascular development, monocyte to macrophage differentiation, bone metabolism, and cancer progression. In this mini-review, we will focus on discussing the emerging roles of QKI in regulating cardiac and vascular development and function and its potential link to cardiovascular pathophysiology.

Global Single Clustering of Phenotype-Associated Human Aging Genes in the Co-Expression and Physical Interaction Networks: An OMIM-Based Investigative Review

BACKGROUND

While a large wealth of literature on aging pertains to in silico, experimental, and predicted genes, many of those genes do not have validated phenotypic consequences in human. Online Mendelian Inheritance in Man (OMIM) provides an exceptional compendium of authoritative, validated aging genes and phenotypes, the interactions among which may enhance the overall perspective of aging mechanisms in human.

METHODS

Here, we reviewed and investigated the global clustering pattern of the OMIM-indexed aging genes (until April 2021) in the gene co-expression and physical interaction networks, using the two keywords "aging" and "ageing". To allow for validity check, we randomly selected six sets of genes from the human genome as control genes, each set consisting of a similar number of genes obtained from the OMIM search. STRING was implemented in the weighted setting and using the edge betweenness parameter, to construct the integrated and tissue-specific networks of the age-related and control genes.

RESULTS

286 aging (ageing) genes and a wide spectrum of 96 associated phenotypes were detected, including late-onset neurodegenerative disorders, cancers, osteoarthritis, and longevity. Despite the general terms used and the vast range of age-related phenotypes, we detected single clustering of the OMIM-extracted aging (ageing) genes in each of the integrated weighted co-expression and physical interaction networks (p<0.0005), as opposed to multiple clustering of the control genes (p≥0.04). TP53 was the overlapping hub gene in each of the networks. Three genes, TP53, APP, and SIRT1 were the consistent hub genes co-expressed across eleven selected human tissues frequently affected by age-related phenotypes.

CONCLUSION

We propose predominant single clustering of the human phenotype-associated aging genes in the co-expression and physical interaction networks, and list the top pathways and genes involved.

Artificial intelligence and machine learning-aided drug discovery in central nervous system diseases: State-of-the-arts and future directions

Neurological disorders significantly outnumber diseases in other therapeutic areas. However, developing drugs for central nervous system (CNS) disorders remains the most challenging area in drug discovery, accompanied with the long timelines and high attrition rates. With the rapid growth of biomedical data enabled by advanced experimental technologies, artificial intelligence (AI) and machine learning (ML) have emerged as an indispensable tool to draw meaningful insights and improve decision making in drug discovery. Thanks to the advancements in AI and ML algorithms, now the AI/ML-driven solutions have an unprecedented potential to accelerate the process of CNS drug discovery with better success rate. In this review, we comprehensively summarize AI/ML-powered pharmaceutical discovery efforts and their implementations in the CNS area. After introducing the AI/ML models as well as the conceptualization and data preparation, we outline the applications of AI/ML technologies to several key procedures in drug discovery, including target identification, compound screening, hit/lead generation and optimization, drug response and synergy prediction, de novo drug design, and drug repurposing. We review the current state-of-the-art of AI/ML-guided CNS drug discovery, focusing on blood-brain barrier permeability prediction and implementation into therapeutic discovery for neurological diseases. Finally, we discuss the major challenges and limitations of current approaches and possible future directions that may provide resolutions to these difficulties.

Novel classifier orthologs of bovine and human oocytes matured in different melatonin environments

It has been demonstrated that melatonin influences the developmental competence of both in vivo and in vitro matured oocytes. It modulates oocyte-specific gene expression patterns among mammalian species. Due to differences among study systems, the identification of the classifier orthologs-the homologous genes related among mammals that could universally categorize oocytes matured in environments with varied melatonin levels is still limitedly studied. To gain insight into such orthologs, cross-species transcription profiling meta-analysis of in vitro matured bovine oocytes and in vivo matured human oocytes in low and high melatonin environments was demonstrated in the current study. RNA-Seq data of bovine and human oocytes were retrieved from the Sequence Read Archive database and pre-processed. The used datasets of bovine oocytes obtained from culturing in the absence of melatonin and human oocytes from old patients were regarded as oocytes in the low melatonin environment (Low). Datasets from bovine oocytes cultured in 10-9 M melatonin and human oocytes from young patients were considered as oocytes in the high melatonin environment (High). Candidate orthologs differentially expressed between Low and High melatonin environments were selected by a linear model, and were further verified by Zero-inflated regression analysis. Support Vector Machine (SVM) was applied to determine the potentials of the verified orthologs as classifiers of melatonin environments. According to the acquired results, linear model analysis identified 284 candidate orthologs differentially expressed between Low and High melatonin environments. Among them, only 15 candidate orthologs were verified by Zero-inflated regression analysis (FDR ≤ 0.05). Utilization of the verified orthologs as classifiers in SVM resulted in the precise classification of oocyte learning datasets according to their melatonin environments (Misclassification rates < 0.18, area under curves > 0.9). In conclusion, the cross-species RNA-Seq meta-analysis to identify novel classifier orthologs of matured oocytes under different melatonin environments was successfully demonstrated in this study-delivering candidate orthologs for future studies at biological levels. Such verified orthologs might provide valuable evidence about melatonin sufficiency in target oocytes-by which, the decision on melatonin supplementation could be implied.