Explainable machine learning model for predicting the occurrence of postoperative malnutrition in children with congenital heart disease

Clinical Nutrition, Journal Year: 2021, Volume and Issue: 41(1), P. 202 - 210

Published: Nov. 10, 2021

Language: Английский

---

AMR-Diag: Neural network based genotype-to-phenotype prediction of resistance towards β-lactams in Escherichia coli and Klebsiella pneumoniae

Computational and Structural Biotechnology Journal, Journal Year: 2021, Volume and Issue: 19, P. 1896 - 1906

Published: Jan. 1, 2021

Antibiotic resistance poses a major threat to public health. More effective ways of the antibiotic prescription are needed delay spread resistance. Employment sequencing technologies coupled with use trained neural network algorithms for genotype-to-phenotype prediction will reduce time susceptibility profile identification from days hours. In this work, we have sequenced and phenotypically characterized 171 clinical isolates Escherichia coli Klebsiella pneumoniae Norway India. Based on data, created networks predict ampicillin, 3rd generation cephalosporins carbapenems. All were unassembled enabling within minutes after information becomes available. Moreover, they can be used both Illumina MinION generated data do not require high genome coverage phenotype prediction. We cross-checked our previously published their corresponding datasets. Besides, also an ensemble different datasets, which improved cross-dataset compared single network. Additionally, direct spiked blood cultures found that AMR-Diag networks, sequencing, bacterial species, resistome, as fast 1–8 h start. To knowledge, is first study prediction: (1) employing method; (2) using more than one platform; (3) utilizing sequence cultures.

Language: Английский

---

A Practical Approach for Predicting Antimicrobial Phenotype Resistance in Staphylococcus aureus Through Machine Learning Analysis of Genome Data

Frontiers in Microbiology, Journal Year: 2022, Volume and Issue: 13

Published: March 2, 2022

With the reduction in sequencing price and acceleration of speed, it is particularly important to directly link genotype phenotype bacteria. Here, we firstly predicted minimum inhibitory concentrations ten antimicrobial agents for Staphylococcus aureus using 466 isolates by extracting k-mer from whole genome data combined with three machine learning algorithms: random forest, support vector machine, XGBoost. Considering one two-fold dilution, essential agreement category could reach >85% >90% most agents. For clindamycin, cefoxitin trimethoprim-sulfamethoxazole, >91% >93%, providing information clinical treatment. The successful prediction resistance showed that model identify methicillin-resistant S. aureus. results suggest small datasets available large hospitals bypass existing basic research known genes accurately predict bacterial phenotype.

Language: Английский

---

Artificial Intelligence in Drug Discovery and Development Against Antimicrobial Resistance: A Narrative Review

Iranian Journal of Medical Microbiology, Journal Year: 2024, Volume and Issue: 18(3), P. 135 - 147

Published: Aug. 18, 2024

Language: Английский

---

Artificial Intelligence and Antibiotic Discovery

Antibiotics, Journal Year: 2021, Volume and Issue: 10(11), P. 1376 - 1376

Published: Nov. 10, 2021

Over recent decades, a new antibiotic crisis has been unfolding due to decreased research in this domain, low return of investment for the companies that developed drug, lengthy and difficult process, success rate candidate molecules, an increased use antibiotics farms overall inappropriate antibiotics. This led series pathogens developing resistance, which poses severe threats public health systems while also driving up costs hospitalization treatment. Moreover, without proper action collaboration between academic institutions, catastrophic trend might develop, with possibility returning pre-antibiotic era. Nevertheless, emerging AI-based technologies have started enter field drug development, offering perspective ever-growing problem. Cheaper faster can be achieved through algorithms identify hit compounds, thereby further accelerating development antibiotics, represents vital step solving current crisis. The aim review is provide extended overview artificial intelligence-based are used discovery, together their technological economic impact on industrial sector.

Language: Английский

---

Development and performance assessment of novel machine learning models to predict pneumonia after liver transplantation

Respiratory Research, Journal Year: 2021, Volume and Issue: 22(1)

Published: March 31, 2021

Abstract Background Pneumonia is the most frequently encountered postoperative pulmonary complications (PPC) after orthotopic liver transplantation (OLT), which cause high morbidity and mortality rates. We aimed to develop a model predict pneumonia in OLT patients using machine learning (ML) methods. Methods Data of 786 adult underwent at Third Affiliated Hospital Sun Yat-sen University from January 2015 September 2019 was retrospectively extracted electronic medical records randomly subdivided into training set testing set. With set, six ML models including logistic regression (LR), support vector (SVM), random forest (RF), adaptive boosting (AdaBoost), extreme gradient (XGBoost) (GBM) were developed. These assessed by area under curve (AUC) receiver operating characteristic on The related risk factors outcomes also probed based chosen model. Results 591 eventually included 253 (42.81%) diagnosed with pneumonia, associated increased hospitalization ( P < 0.05). Among models, XGBoost performed best. AUC 0.734 (sensitivity: 52.6%; specificity: 77.5%). notably 14 items features: INR, HCT, PLT, ALB, ALT, FIB, WBC, PT, serum Na + , TBIL, anesthesia time, preoperative length stay, total fluid transfusion operation time. Conclusion Our study firstly demonstrated that common variables might patients.

Language: Английский

---

Machine learning and feature extraction for rapid antimicrobial resistance prediction of Acinetobacter baumannii from whole-genome sequencing data

Frontiers in Microbiology, Journal Year: 2024, Volume and Issue: 14

Published: Jan. 11, 2024

Whole-genome sequencing (WGS) has contributed significantly to advancements in machine learning methods for predicting antimicrobial resistance (AMR). However, the comparisons of different AMR prediction without requiring prior knowledge remains be conducted.

Language: Английский

---

Antimicrobial resistance and machine learning: past, present, and future

Frontiers in Microbiology, Journal Year: 2023, Volume and Issue: 14

Published: May 26, 2023

Machine learning has become ubiquitous across all industries, including the relatively new application of predicting antimicrobial resistance. As first bibliometric review in this field, we expect it to inspire further research area. The employs standard indicators such as article count, citation and Hirsch index (H-index) evaluate relevance impact leading countries, organizations, journals, authors field. VOSviewer Biblioshiny programs are utilized analyze co-citation networks, collaboration keyword co-occurrence, trend analysis. United States highest contribution with 254 articles, accounting for over 37.57% total corpus, followed by China (103) Kingdom (78). Among 58 publishers, top four publishers account 45% publications, Elsevier 15% Springer Nature (12%), MDPI, Frontiers Media SA 9% each. Microbiology is most frequent publication source (33 articles), Scientific Reports (29 PLoS One (17 Antibiotics (16 articles). study reveals a substantial increase publications on use machine predict antibiotic Recent focused developing advanced algorithms that can accurately forecast resistance, range now being used address issue.

Language: Английский

---

Accessory genes define species-specific routes to antibiotic resistance

Life Science Alliance, Journal Year: 2024, Volume and Issue: 7(4), P. e202302420 - e202302420

Published: Jan. 16, 2024

A deeper understanding of the relationship between antimicrobial resistance (AMR) gene carriage and phenotype is necessary to develop effective response strategies against this global burden. AMR often a result multi-gene interactions; therefore, we need approaches that go beyond current simple identification tools. Machine-learning (ML) methods may meet challenge allow development rapid computational for classification. To examine this, applied multiple ML techniques 16,950 bacterial genomes across 28 genera, with corresponding MICs 23 antibiotics aim training models accurately determine from sequenced genomes. This resulted in >1.5-fold increase prediction accuracy over alone. Furthermore, revealed 528 unique (often species-specific) genomic routes antibiotic resistance, including genes not previously linked phenotype. Our study demonstrates utility predicting phenotypes diverse clinically relevant organisms antibiotics. research proposes method support laboratory-based pathogens.

Language: Английский

---

Characterization of immune microenvironment associated with medulloblastoma metastasis based on explainable machine learning

Pediatric Investigation, Journal Year: 2025, Volume and Issue: 9(1), P. 59 - 69

Published: Feb. 14, 2025

ABSTRACT Importance Medulloblastoma (MB) is the most common malignant brain tumor in children, with metastasis being primary cause of recurrence and mortality. The microenvironment (TME) plays a critical role driving metastasis; however, mechanisms underlying TME alterations MB remain poorly understood. Objective To develop validate machine learning (ML) models for predicting patient outcomes to investigate components, particularly immune cells immunoregulatory molecules, metastasis. Methods ML were constructed validated predict prognosis patients. Eight algorithms evaluated, optimal model was selected. Lasso regression employed feature selection, SHapley Additive exPlanations values used interpret contribution individual features predictions. Immune cell infiltration tissues quantified using populations‐counter method, immunohistochemistry applied analyze expression distribution specific proteins tissues. Results identified as strongest predictor poor patients, significantly worse survival observed metastatic cases. High CD8+ T cytotoxic lymphocytes (CTLs), along elevated TGFB1 gene encoding transforming growth factor beta 1 (TGF‐β1), strongly associated Independent transcriptomic immunohistochemical analyses confirmed higher cell/CTL TGF‐β1 compared nonmetastatic samples. Patients both high context exhibited patients low no Interpretation This study identifies key prognostic reveals pivotal roles cells, CTLs, within promoting outcomes. These findings provide foundation developing future therapeutic strategies targeting improve

Language: Английский

---