Fully Automated Unconstrained Analysis of High-Resolution Mass Spectrometry Data with Machine Learning

Journal of the American Chemical Society, Journal Year: 2022, Volume and Issue: 144(32), P. 14590 - 14606

Published: Aug. 8, 2022

Mass spectrometry (MS) is a convenient, highly sensitive, and reliable method for the analysis of complex mixtures, which vital materials science, life sciences fields such as metabolomics proteomics, mechanistic research in chemistry. Although it one most powerful methods individual compound detection, complete signal assignment mixtures still great challenge. The unconstrained formula-generating algorithm, covering entire spectra revealing components, "dream tool" researchers. We present framework efficient MS data interpretation, describing novel approach detailed based on deisotoping performed by gradient-boosted decision trees neural network that generates molecular formulas from fine isotopic structure, approaching long-standing inverse spectral problem. were successfully tested three examples: fragment ion protein sequencing natural samples sciences, study cross-coupling catalytic system

Language: Английский

---

Single-cell metabolomics: where are we and where are we going?

Current Opinion in Biotechnology, Journal Year: 2022, Volume and Issue: 75, P. 102693 - 102693

Published: Feb. 10, 2022

Single-cell metabolomics with mass spectrometry enables a large variety of metabolites to be simultaneously detected from individual cells, without any preselection or labelling, map phenotypes on the single cell level. Although field is relatively young, it steadily progressing an increasing number active research groups, techniques for sampling and ionization, tools data analysis, applications answer important biomedical environmental questions. In addition, community shows great creativity in overcoming challenges associated low sample volumes, wide range metabolite species, datasets. Here, we briefly discuss publications since 2019 aim provide unfamiliar reader insight into expert update current status field.

Language: Английский

---

Deep learning in spectral analysis: Modeling and imaging

TrAC Trends in Analytical Chemistry, Journal Year: 2024, Volume and Issue: 172, P. 117612 - 117612

Published: Feb. 20, 2024

Language: Английский

---

Secondary ion mass spectrometry

Nature Reviews Methods Primers, Journal Year: 2024, Volume and Issue: 4(1)

Published: May 9, 2024

Language: Английский

---

Image-guided MALDI mass spectrometry for high-throughput single-organelle characterization

Nature Methods, Journal Year: 2021, Volume and Issue: 18(10), P. 1233 - 1238

Published: Sept. 30, 2021

Abstract Peptidergic dense-core vesicles are involved in packaging and releasing neuropeptides peptide hormones—critical processes underlying brain, endocrine exocrine function. Yet, the heterogeneity within these organelles, even for morphologically defined vesicle types, is not well characterized because of their small volumes. We present image-guided, high-throughput mass spectrometry-based protocols to chemically profile large populations both lucent lipid contents, allowing observation chemical between two populations. The proteolytic processing products four prohormones observed vesicles, spectral features corresponding specific suggest three distinct Notable differences range vesicles. These single-organelle spectrometry approaches adaptable characterize a subcellular structures.

Language: Английский

---

Interpretable heartbeat classification using local model-agnostic explanations on ECGs

Computers in Biology and Medicine, Journal Year: 2021, Volume and Issue: 133, P. 104393 - 104393

Published: April 16, 2021

Language: Английский

---

A Review on Interpretable and Explainable Artificial Intelligence in Hydroclimatic Applications

Water, Journal Year: 2022, Volume and Issue: 14(8), P. 1230 - 1230

Published: April 11, 2022

This review focuses on the use of Interpretable Artificial Intelligence (IAI) and eXplainable (XAI) models for data imputations numerical or categorical hydroclimatic predictions from nonlinearly combined multidimensional predictors. The AI considered in this paper involve Extreme Gradient Boosting, Light Categorical Extremely Randomized Trees, Random Forest. These can transform into XAI when they are coupled with explanatory methods such as Shapley additive explanations local interpretable model-agnostic explanations. highlights that IAI capable unveiling rationale behind while discovering new knowledge justifying AI-based results, which critical enhanced accountability AI-driven predictions. also elaborates importance domain interventional modeling, potential advantages disadvantages hybrid non-IAI predictive unequivocal balanced decisions, choice performance versus physics-based modeling. concludes a proposed framework to enhance interpretability explainability applications.

Language: Английский

---

Advanced mass spectrometric and spectroscopic methods coupled with machine learning for in vitro diagnosis

View, Journal Year: 2022, Volume and Issue: 4(1)

Published: Dec. 28, 2022

Abstract In vitro diagnosis (IVD) is one vital component of medical tests that detects biological samples tissues or bio‐fluids. Recently, mass spectrometry and spectroscopy have been increasingly employed in the field IVD, due to their high accuracy, facile sample preparation, rapid detection. Notably, large datasets generated by these two technology methods provide a wealth information but subsequently involve complex time‐consuming processing works. Machine learning (ML), an important branch artificial intelligence (AI), has emerged as promising solution for decoding big data. ML imitates human brain process data, significantly improving accuracy efficiency compared with traditional methods. this review, we first introduce commonly used algorithms advanced techniques respectively. The are summarized four aspects according different tasks. Then, combinations spectrometry, spectroscopy, multi‐modal analysis IVD presented, roles elucidated some representative examples. This review aims systematic comprehensive summary literature on ML‐assisted spectroscopy. We believe it will facilitate researchers select suitable supplementing existing detection develop potential coupling more ML, thus promoting development IVD.

Language: Английский

---

AI/ML-driven advances in untargeted metabolomics and exposomics for biomedical applications

Cell Reports Physical Science, Journal Year: 2022, Volume and Issue: 3(7), P. 100978 - 100978

Published: July 1, 2022

Metabolomics describes a high-throughput approach for measuring repertoire of metabolites and small molecules in biological samples. One utility untargeted metabolomics, unbiased global analysis the metabolome, is to detect key as contributors to, or readouts of, human health disease. In this perspective, we discuss how artificial intelligence (AI) machine learning (ML) have promoted major advances metabolomics workflows facilitated pivotal findings areas disease screening diagnosis. We contextualize applications AI ML emerging field high-resolution mass spectrometry (HRMS) exposomics, which unbiasedly detects endogenous exogenous chemicals tissue characterize exposure linked with outcomes. state science suggest potential opportunities using improve data quality, rigor, detection, chemical identification exposomics studies.

Language: Английский

---

Interpretable machine learning with tree-based shapley additive explanations: Application to metabolomics datasets for binary classification

PLoS ONE, Journal Year: 2023, Volume and Issue: 18(5), P. e0284315 - e0284315

Published: May 4, 2023

Machine learning (ML) models are used in clinical metabolomics studies most notably for biomarker discoveries, to identify metabolites that discriminate between a case and control group. To improve understanding of the underlying biomedical problem bolster confidence these model interpretability is germane. In metabolomics, partial least square discriminant analysis (PLS-DA) its variants widely used, partly due model's with Variable Influence Projection (VIP) scores, global interpretable method. Herein, Tree-based Shapley Additive explanations (SHAP), an ML method grounded game theory, was explain local explanation properties. this study, experiments (binary classification) were conducted three published datasets using PLS-DA, random forests, gradient boosting, extreme boosting (XGBoost). Using one datasets, PLS-DA explained VIP while best-performing models, forest model, interpreted Tree SHAP. The results show SHAP has more depth than PLS-DA's VIP, making it powerful rationalizing machine predictions from studies.

Language: Английский

---