DeepThal: A Deep Learning-Based Framework for the Large-Scale Prediction of the α+-Thalassemia Trait Using Red Blood Cell Parameters

Journal of Clinical Medicine, Journal Year: 2022, Volume and Issue: 11(21), P. 6305 - 6305

Published: Oct. 26, 2022

Objectives: To develop a machine learning (ML)-based framework using red blood cell (RBC) parameters for the prediction of α+-thalassemia trait (α+-thal trait) and to compare diagnostic performance with conventional method single RBC parameter or combination parameters. Methods: A retrospective study was conducted on possible couples at risk fetus hemoglobin H (Hb disease). Subjects molecularly confirmed normal status (not thalassemia), α+-thal trait, two-allele α-thalassemia mutation were included. Clinical (age gender) (Hb, Hct, MCV, MCH, MCHC, RDW, count) obtained from their antenatal thalassemia screen retrieved analyzed method. The evaluated. Results: In total, 594 cases (female/male: 330/264, mean age: 29.7 ± 6.6 years) included in analysis. There 229 controls, 160 205 category, respectively. ML-derived model improved performance, giving sensitivity 80% specificity 81%. experimental results indicated that DeepThal achieved better compared other ML-based methods terms independent test dataset, an accuracy 80.77%, 70.59%, Matthews correlation coefficient (MCC) 0.608. Of all parameters, MCH < 28.95 pg as had highest predicting AUC 0.857 95% CI 0.816−0.899. derived binary logistic regression analysis exhibited 0.868 0.830−0.906, 80.1% 75.1%. Conclusions: dataset is sufficient demonstrate capable accurately trait. It anticipated will be useful tool scientific community large-scale

Language: Английский

---

Pretoria: An effective computational approach for accurate and high-throughput identification of CD8+ t-cell epitopes of eukaryotic pathogens

International Journal of Biological Macromolecules, Journal Year: 2023, Volume and Issue: 238, P. 124228 - 124228

Published: March 29, 2023

Language: Английский

---

PINNED: identifying characteristics of druggable human proteins using an interpretable neural network

Journal of Cheminformatics, Journal Year: 2023, Volume and Issue: 15(1)

Published: July 19, 2023

Abstract The identification of human proteins that are amenable to pharmacologic modulation without significant off-target effects remains an important unsolved challenge. Computational methods have been devised identify features which distinguish between “druggable” and “undruggable” proteins, finding protein sequence, tissue cellular localization, biological role, position in the protein–protein interaction network all discriminant factors. However, many prior efforts automate assessment druggability suffer from low performance or poor interpretability. We developed a neural network-based machine learning model capable generating sub-scores based on each four distinct categories, combining them form overall score. achieves excellent separating drugged undrugged proteome, with area under receiver operating characteristic (AUC) 0.95. Our use multiple allows potential targets interest contributors druggability, leading more interpretable holistic novel targets.

Language: Английский

---

StackTTCA: a stacking ensemble learning-based framework for accurate and high-throughput identification of tumor T cell antigens

BMC Bioinformatics, Journal Year: 2023, Volume and Issue: 24(1)

Published: July 28, 2023

Abstract Background The identification of tumor T cell antigens (TTCAs) is crucial for providing insights into their functional mechanisms and utilizing potential in anticancer vaccines development. In this context, TTCAs are highly promising. Meanwhile, experimental technologies discovering characterizing new expensive time-consuming. Although many machine learning (ML)-based models have been proposed identifying TTCAs, there still a need to develop robust model that can achieve higher rates accuracy precision. Results study, we propose stacking ensemble learning-based framework, termed StackTTCA, accurate large-scale TTCAs. Firstly, constructed 156 different baseline by using 12 feature encoding schemes 13 popular ML algorithms. Secondly, these were trained employed create probabilistic vector. Finally, the optimal vector was determined based selection strategy then used construction our stacked model. Comparative benchmarking experiments indicated StackTTCA clearly outperformed several classifiers existing methods terms independent test, with an 0.932 Matthew's correlation coefficient 0.866. Conclusions summary, framework could help precisely rapidly identify true follow-up verification. addition, developed online web server ( http://2pmlab.camt.cmu.ac.th/StackTTCA ) maximize user convenience high-throughput screening novel

Language: Английский

---

TROLLOPE: A novel sequence-based stacked approach for the accelerated discovery of linear T-cell epitopes of hepatitis C virus

PLoS ONE, Journal Year: 2023, Volume and Issue: 18(8), P. e0290538 - e0290538

Published: Aug. 25, 2023

Hepatitis C virus (HCV) infection is a concerning health issue that causes chronic liver diseases. Despite many successful therapeutic outcomes, no effective HCV vaccines are currently available. Focusing on T cell activity, the primary effector for clearance, epitopes of (TCE-HCV) considered promising elements to accelerate vaccine efficacy. Thus, accurate and rapid identification TCE-HCVs recommended obtain more efficient therapy infection. In this study, novel sequence-based stacked approach, termed TROLLOPE, proposed accurately identify from sequence information. Specifically, we employed 12 different feature descriptors heterogeneous perspectives, such as physicochemical properties, composition-transition-distribution information composition These were used in cooperation with popular machine learning (ML) algorithms create 144 base-classifiers. To maximize utility these base-classifiers, selection strategy determine collection potential base-classifiers integrated them develop meta-classifier. Comprehensive experiments based both cross-validation independent tests demonstrated superior predictive performance TROLLOPE compared conventional ML classifiers, test accuracies 0.745 0.747, respectively. Finally, user-friendly online web server ( http://pmlabqsar.pythonanywhere.com/TROLLOPE ) has been developed serve research efforts large-scale follow-up experimental verification.

Language: Английский

---

Druggable protein prediction using a multi-canal deep convolutional neural network based on autocovariance method

Computers in Biology and Medicine, Journal Year: 2022, Volume and Issue: 151, P. 106276 - 106276

Published: Nov. 8, 2022

Language: Английский

---

Deciphering Breast Cancer Metastasis Cascade: A Systems Biology Approach Integrating Transcriptome and Interactome Insights for Target Discovery

OMICS A Journal of Integrative Biology, Journal Year: 2024, Volume and Issue: 28(3), P. 148 - 161

Published: March 1, 2024

Breast cancer is the lead cause of cancer-related deaths among women globally. metastasis a complex and still inadequately understood process key dimension mortality attendant to breast cancer. This study reports dysregulated genes across metastatic stages tissues, shedding light on their molecular interplay in disease pathogenesis new possibilities for drug discovery. Comprehensive analyses gene expression data from primary tumor, circulating tumor cells, distant sites brain, lung, liver, bone were conducted. Genes multiple tissues identified as cascade genes, are further classified based functional associations with metastasis-related mechanisms. Their interactions HUB interactome networks scrutinized, followed by pathway enrichment analysis. Validation potential targets included assessments survival, druggability, prognostic marker status, secretome annotation, protein expression, cell type association. Results displayed critical those specific sites, revealing involvement collagen degradation assembly fibrils other multimeric structure pathways driving metastasis. Notably, pivotal FABP4, CXCL12, APOD, IGF1 emerged high potential, linked significant druggability survival scores, establishing them targets. The significance this research lies its uncover novel biomarkers early detection, therapeutic targets, deeper understanding mechanisms underpinning cancer, an eye precision/personalized medicine.

Language: Английский

---

Empirical comparison and analysis of machine learning-based approaches for druggable protein identification.

PubMed, Journal Year: 2023, Volume and Issue: 22, P. 915 - 927

Published: Jan. 1, 2023

Efficiently and precisely identifying drug targets is crucial for developing discovering potential medications. While conventional experimental approaches can accurately pinpoint these targets, they suffer from time constraints are not easily adaptable to high-throughput processes. On the other hand, computational approaches, particularly those utilizing machine learning (ML), offer an efficient means accelerate prediction of druggable proteins based solely on their primary sequences. Recently, several state-of-the-art methods have been developed predicting analyzing proteins. These showed high diversity in terms benchmark datasets, feature extraction schemes, ML algorithms, evaluation strategies webserver/software usability. Thus, our objective reexamine conduct a comprehensive assessment strengths weaknesses across multiple aspects. In this study, we deliver first survey regarding silico First, provided information existing datasets types employed. Second, investigated effectiveness protein identification each dataset. Third, summarized important features used field webserver/software. Finally, addressed present valuable guidance scientific community designing novel models. We anticipate that review will provide development more accurate predictors.

Language: Английский

---

The usability of stacking-based ensemble learning model in crime prediction: a systematic review

Crime Prevention and Community Safety, Journal Year: 2024, Volume and Issue: 26(4), P. 440 - 489

Published: Nov. 20, 2024

This research addresses the potential for tackling crime volumes and improving analytics through new enhancement strategies. The use of machine learning deep solutions is increasing in prediction, as many other fields. study aims to strengthen proactive approaches criminology by evaluating effectiveness stacking-based ensemble (S-BEL) model, which enhance overall performance combining strengths various algorithms improve facilitate prevention analyzes six studies leveraging S-BEL model along with 28 articles on seven utilizing models, 56 general prediction studies. findings highlight that stands out a prominent technique providing valuable insights law enforcement.

Language: Английский

---

Repurposing FDA-Approved Drugs Against Potential Drug Targets Involved in Brain Inflammation Contributing to Alzheimer’s Disease

Targets, Journal Year: 2024, Volume and Issue: 2(4), P. 446 - 469

Published: Dec. 4, 2024

Alzheimer's disease is a neurodegenerative that continues to have rising number of cases. While extensive research has been conducted in the last few decades, only drugs approved by FDA for treatment, and even fewer aim be curative rather than manage symptoms. There remains an urgent need understanding pathogenesis, as well identifying new targets further drug discovery. (AD) known stem from build-up amyloid beta (Aβ) plaques tangles tau proteins. Furthermore, inflammation brain arise degeneration tissue insoluble material. Therefore, there potential link between pathology AD brain, especially progresses later stages where neuronal death levels are higher. Proteins relevant both thus make ideal therapeutics; however, proteins evaluated determine which would therapeutic treatments, or 'druggable'. Druggability analysis was using two structure-based methods (i.e., Drug-Like Density SiteMap), sequence-based approach, SPIDER. The most druggable were then single-nuclei sequencing data their clinical relevance AD. For each top five targets, small molecule docking used evaluate able bind with chosen included DRD2 (inhibits adenylyl cyclase activity), C9 (binds C5B8 form membrane attack complex), C4b C2a C3 convertase), C5AR1 (GPCR binds C5a), GABA-A-R involved inhibiting neurotransmission). Each target had multiple inhibitors FDA-approved list decent binding infinities. Among these inhibitors, found more one protein target. They C15H14N2O2 v316 (Paracetamol), treat pain/inflammation originally cataracts relieve headaches/fever, respectively. These results provide groundwork experimental investigation trials.

Language: Английский

---