Limits and Prospects of Molecular Fingerprinting for Phenotyping Biological Systems Revealed through <i>In Silico</i> Modeling

Eissa, Tarek; Kepesidis, Kosmas V.; Žigman, Mihaela; Huber, Maximilian

doi:10.1021/acs.analchem.2c04711

Cited by 8 publications

(3 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The concept of within-person stability of IR fingerprints has been revealed, 78 and we also demonstrated the underlying concept that time-tracking of IR fingerprints enhances phenotype detection. 79 However, it still remains to be proven whether decades-long health state trajectories can be decoded with IR fingerprinting.…”

Section: Discussionmentioning

confidence: 99%

Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening

Eissa,

Leonardo,

Kepesidis

et al. 2024

Cell Reports Medicine

Self Cite

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening

Eissa,

Leonardo,

Kepesidis

et al. 2024

Cell Reports Medicine

Self Cite

View full text Add to dashboard Cite

“…2a, upper left), as defined previously [29]. Alternatively, opting to set the calibration measurements to be of different individuals would yield a level of between-person biological variability, as demonstrated previously [31].…”

Section: Characterizing Empirically-observed Variabilitymentioning

confidence: 95%

“…We previously introduced an in silico model that generates 1D spectra of complex biological samples, focusing on IR absorption spectra [31]. Our initial work explored the impact of varying levels of between-person biological variability on classification efficiency in simulated case-control conditions.…”

Section: Characterizing Empirically-observed Variabilitymentioning

confidence: 99%

CODI: Enhancing machine learning-based molecular profiling through contextual out-of-distribution integration

Eissa,

Huber,

Obermayer-Pietsch

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Molecular analytics increasingly utilize machine learning (ML) for predictive modeling based on data acquired through molecular profiling technologies. However, developing robust models that accurately capture physiological phenotypes is challenged by a multitude of factors. These include the dynamics inherent to biological systems, variability stemming from analytical procedures, and the resource-intensive nature of obtaining sufficiently representative datasets. Here, we propose and evaluate a new method: Contextual Out-of-Distribution Integration (CODI). Based on experimental observations, CODI generates synthetic data that integrate unrepresented sources of variation encountered in real-world applications into a given molecular fingerprint dataset. By augmenting a dataset with out-of-distribution variance, CODI enables an ML model to better generalize to samples beyond the initial training data. Using three independent longitudinal clinical studies and a case-control study, we demonstrate CODI’s application to several classification scenarios involving vibrational spectroscopy of human blood. We showcase our approach’s ability to enable personalized fingerprinting for multi-year longitudinal molecular monitoring and enhance the robustness of trained ML models for improved disease detection. Our comparative analyses revealed that incorporating CODI into the classification workflow consistently led to significantly improved classification accuracy while minimizing the requirement of collecting extensive experimental observations.SIGNIFICANCE STATEMENTAnalyzing molecular fingerprint data is challenging due to multiple sources of biological and analytical variability. This variability hinders the capacity to collect sufficiently large and representative datasets that encompass realistic data distributions. Consequently, the development of machine learning models that generalize to unseen, independently collected samples is often compromised. Here, we introduce CODI, a versatile framework that enhances traditional classifier training methodologies. CODI is a general framework that incorporates information about possible out-of-distribution variations into a given training dataset, augmenting it with simulated samples that better capture the true distribution of the data. This allows the classification to achieve improved predictive performance on samples beyond the original distribution of the training data.

show abstract

Die Herausforderungen der molekularen Interpretationen von Infrarotspektren komplexer Proben

Eissa,

Voronina,

Huber

et al. 2024

Angewandte Chemie

View full text Add to dashboard Cite

Die Schwingungsspektroskopie ist eine weit verbreitete Technik zur chemischen Charakterisierung in verschiedenen analytischen Disziplinen. Ihre Anwendungen erstrecken sich zunehmend auf die Analyse komplexer Proben wie Bioflüssigkeiten und ermöglichen molekulares Profiling mit hohem Durchsatz. Trotz ihrer Leistungsfähigkeit leidet diese Technologie unter einer inhärenten Einschränkung: Der Überlapp von Absorptionsinformationen über verschiedene Spektralbereiche hinweg behindert die Fähigkeit, einzelne Substanzen zu identifizieren, welche zu den gemessenen Gesamtsignal beitragen. Obwohl man sich dieser Herausforderung bewusst ist, wird die Schwierigkeit der Analyse von Multimolekülspektren oft unterschätzt, was zu Fehlinterpretation führen kann. In dieser Arbeit diskutieren wir kritisch den weit verbreiteten übermäßigen Verlass auf einzelne Absorptionsbanden bei der Dateninterpretation und beleuchten die Fallstricke bei der Korrelation von spektral Signalen mit diskreten Substanzen oder physiologischen Zuständen ohne rigorose Validierung. Mit Fokus auf blutbasierte Infrarotspektroskopie liefern wir Beispiele, die zeigen, wie Überlappungen lokaler Absorptionsmaxima zwischen verschiedenen Substanzen, relative Konzentrationen von Substanzen und Datenvorverarbeitungsschritte zu fehlerhaften Interpretationen führen können. Wir plädieren für einen Paradigmenwechsel hin zu einem vorsichtigeren Verständnis komplexer Spektren, was dazu führen sollte, entweder deren Charakter als molekularen Fingerabdruck zu akzeptieren und maschinelles Lernen zur Analyse zu nutzen – oder zusätzliche Messmodalitäten für eine robuste molekulare Interpretationen einzubeziehen. Mit dem Ziel, analytische Praktiken in diesem Bereich zu verbessern und weiterzuentwickeln, heben wir die Grenzen molekularer Interpretationen hervor und stellen mögliche Anwendungen vor.

show abstract

Limits and Prospects of Molecular Fingerprinting for Phenotyping Biological Systems Revealed through In Silico Modeling

Cited by 8 publications

References 40 publications

Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening

Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening

CODI: Enhancing machine learning-based molecular profiling through contextual out-of-distribution integration

Die Herausforderungen der molekularen Interpretationen von Infrarotspektren komplexer Proben

Contact Info

Product

Resources

About