Vegetable waste and by-products to feed a healthy gut microbiota: Current evidence, machine learning and computational tools to design novel microbiome-targeted foods

Trends in Food Science & Technology, Journal Year: 2021, Volume and Issue: 118, P. 399 - 417

Published: Oct. 7, 2021

Food waste management is a key issue to global food security and friendly environmental governance. Worldwide, one-third of produced for human consumption lost or wasted along the supply chain, primary production processing representing most significant loses. Therefore, need achieve zero schemes becoming priority meet Sustainable Development Goals. Increasing evidence points towards vegetable as rich source wide array carbohydrate structures fibres providing opportunity identify develop alternative approaches valorize agro-food waste. This review describes valorization by-products via (novel) substrates targeted gut microbiota modulation, emphasizing importance raw materials structural-functional properties carbohydrates. Furthermore, we propose novel framework rational selection sources with potential prebiotic activity, based on machine learning other computational tools applied available literature public database information. Integration body knowledge within field valorization, from different perspectives, allows carbohydrate-based promising activities. By exploring interactions among dietary fibre microbial ecosystems using fed structural, functional genomic data, can selectively stimulate commensals, in agreement experimental evidence. Our approach establishes new that be extended range commensal microbes structures.

Language: Английский

---

Microbiome-based disease prediction with multimodal variational information bottlenecks

PLoS Computational Biology, Journal Year: 2022, Volume and Issue: 18(4), P. e1010050 - e1010050

Published: April 11, 2022

Scientific research is shedding light on the interaction of gut microbiome with human host and its role in health. Existing machine learning methods have shown great potential discriminating healthy from diseased states. Most them leverage shotgun metagenomic sequencing to extract microbial species-relative abundances or strain-level markers. Each these profiling modalities showed diagnostic when tested separately; however, no existing approach combines a single predictive framework. Here, we propose Multimodal Variational Information Bottleneck (MVIB), novel deep model capable joint representation multiple heterogeneous data modalities. MVIB achieves competitive classification performance while being faster than methods. Additionally, offers interpretable results. Our adopts an information theoretic interpretation neural networks computes stochastic encoding different input We use predict whether hosts are affected by certain disease jointly analysing evaluated samples 11 publicly available cohorts covering 6 diseases. achieve high (0.80 < ROC AUC 0.95) 5 at least medium remaining ones. adopt saliency technique interpret output identify most relevant species markers model's predictions. also perform cross-study generalisation experiments, where train test same disease, overall comparable results baseline approach, i.e. Random Forest. Further, evaluate our adding metabolomic derived mass spectrometry as third modality. method scalable respect has average training time 1.4 seconds. The source code datasets used this work available.

Language: Английский

---

Artificial intelligence approaches to human-microbiome protein–protein interactions

Current Opinion in Structural Biology, Journal Year: 2022, Volume and Issue: 73, P. 102328 - 102328

Published: Feb. 10, 2022

Host-microbiome interactions play significant roles in human health and disease. Artificial intelligence approaches have been developed to better understand predict the molecular interplay between host its microbiome. Here, we review recent advancements computational methods microbial effects on cells with a special focus protein-protein interactions. We categorize from traditional ones more deep learning methods, followed by several challenges potential solutions structure-based approaches. This serves as brief guide current status future directions field.

Language: Английский

---

Advancing microbiome research with machine learning: key findings from the ML4Microbiome COST action

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

Published: Sept. 21, 2023

The rapid development of machine learning (ML) techniques has opened up the data-dense field microbiome research for novel therapeutic, diagnostic, and prognostic applications targeting a wide range disorders, which could substantially improve healthcare practices in era precision medicine. However, several challenges must be addressed to exploit benefits ML this fully. In particular, there is need establish “gold standard” protocols conducting analysis experiments interactions between researchers experts. Machine Learning Techniques Human Microbiome Studies (ML4Microbiome) COST Action CA18131 European network established 2019 promote collaboration discovery-oriented data-driven experts optimize standardize approaches analysis. This perspective paper presents key achievements ML4Microbiome, include identifying predictive discriminatory ‘omics’ features, improving repeatability comparability, developing automation procedures, defining priority areas methods microbiome. insights gained from ML4Microbiome will help maximize potential pave way new improved practices.

Language: Английский

---

Sequence-Based Prediction of Plant Allergenic Proteins: Machine Learning Classification Approach

ACS Omega, Journal Year: 2023, Volume and Issue: 8(4), P. 3698 - 3704

Published: Jan. 20, 2023

This Article proposes a novel chemometric approach to understanding and exploring the allergenic nature of food proteins. Using machine learning methods (supervised unsupervised), this work aims predict allergenicity plant The strategy is based on scoring descriptors testing their classification performance. Partitioning was support vector machines (SVM), k-nearest neighbor (KNN) classifier applied. A fivefold cross-validation used validate KNN in variable selection step as well final classifier. To overcome problem allergies, robust efficient method for protein needed.

Language: Английский

---

Explainable artificial intelligence and machine learning: novel approaches to face infectious diseases challenges

Annals of Medicine, Journal Year: 2023, Volume and Issue: 55(2)

Published: Nov. 27, 2023

Artificial intelligence (AI) and machine learning (ML) are revolutionizing human activities in various fields, with medicine infectious diseases being not exempt from their rapid exponential growth. Furthermore, the field of explainable AI ML has gained particular relevance is attracting increasing interest. Infectious have already started to benefit AI/ML models. For example, they been employed or proposed better understand complex models aimed at improving diagnosis management coronavirus disease 2019, antimicrobial resistance prediction quantum vaccine algorithms. Although some issues concerning dichotomy between explainability interpretability still require careful attention, an in-depth understanding how arrive predictions recommendations becoming increasingly essential properly face growing challenges present century.

Language: Английский

---

A new era in healthcare: The integration of artificial intelligence and microbial

Medicine in Novel Technology and Devices, Journal Year: 2024, Volume and Issue: 23, P. 100319 - 100319

Published: July 2, 2024

The convergence of artificial intelligence (AI) and microbial therapeutics offers promising avenues for novel discoveries therapeutic interventions. With the exponential growth omics datasets rapid advancements in AI technology, next generation is increasingly prevalent microbiology research. In research, instrumental classification functional annotation microorganisms. Machine learning algorithms facilitate efficient accurate categorization taxa, enabling identification traits metabolic pathways within communities. Additionally, AI-driven protein design strategies hold promise engineering enzymes with enhanced catalytic activities stabilities. By predicting structures, functions, interactions, enable rational proteins tailored specific applications. systems are already present clinical laboratories form expert rules used by some automated susceptibility testing systems. future, technologists will rely more heavily on initial screening, allowing them to focus diagnostic challenges complex technical interpretations. approaches immense advancing our understanding ecosystems, accelerating drug discovery processes, fostering development groundbreaking This review aims summarize common their applications synthetic biology. We provide a comprehensive evaluation AI's utility discussing both its advantages challenges. Finally, we explore future research directions bottlenecks faced field.

Language: Английский

---

Could Artificial Intelligence/Machine Learning and Inclusion of Diet-Gut Microbiome Interactions Improve Disease Risk Prediction? Case Study: Coronary Artery Disease

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

Published: April 11, 2022

Coronary artery disease (CAD) is the most common cardiovascular (CVD) and main leading cause of morbidity mortality worldwide, posing a huge socio-economic burden to society health systems. Therefore, timely precise identification people at high risk CAD urgently required. Most current prediction approaches are based on small number traditional factors (age, sex, diabetes, LDL HDL cholesterol, smoking, systolic blood pressure) incompletely predictive across all patient groups, as multi-factorial with complex etiology, considered be driven by both genetic, well numerous environmental/lifestyle factors. Diet one modifiable for improving lifestyle prevention. However, rise in obesity, type 2 diabetes (T2D) CVD/CAD indicates that “one-size-fits-all” approach may not efficient, due significant variation inter-individual responses. Recently, gut microbiome has emerged potential previously under-explored contributor these variations. Hence, efficient integration dietary information alongside genetic variations clinical data holds great promise improve prediction. Nevertheless, highly nature meals combined variability poses several Big Data analytics challenges modeling diet-gut microbiota interactions integrating within development personalized decision support systems (DSS). In this regard, recent re-emergence Artificial Intelligence (AI) / Machine Learning (ML) opening intriguing perspectives, able capture large matrices data, incorporating their identifying linear non-linear relationships. Mini-Review, we consider (1) used AI/ML different use cases (2) content, choice impact risk; (3) classification individuals composition into vs. controls (4) risk. Finally, provide an outlook putting it together improved predictions.

Language: Английский

---

Overview of data preprocessing for machine learning applications in human microbiome research

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

Published: Oct. 5, 2023

Although metagenomic sequencing is now the preferred technique to study microbiome-host interactions, analyzing and interpreting microbiome data presents challenges primarily attributed statistical specificities of (e.g., sparse, over-dispersed, compositional, inter-variable dependency). This mini review explores preprocessing transformation methods applied in recent human studies address analysis challenges. Our results indicate a limited adoption targeting characteristics data. Instead, there prevalent usage relative normalization-based transformations that do not specifically account for specific attributes The information on before was incomplete or missing many publications, leading reproducibility concerns, comparability issues, questionable results. We hope this will provide researchers newcomers field research with an up-to-date point reference various tools assist them choosing most suitable method based their questions, objectives, characteristics.

Language: Английский

---

AI in microbiome‐related healthcare

Microbial Biotechnology, Journal Year: 2024, Volume and Issue: 17(11)

Published: Nov. 1, 2024

Abstract Artificial intelligence (AI) has the potential to transform clinical practice and healthcare. Following impressive advancements in fields such as computer vision medical imaging, AI is poised drive changes microbiome‐based healthcare while facing challenges specific field. This review describes state‐of‐the‐art use of microbiome‐related It points out limitations across topics data handling, modelling safeguarding patient privacy. Furthermore, we indicate how these current shortcomings could be overcome future discuss influence opportunities increasingly complex on

Language: Английский

---