Logo do repositório

Percorrer por autor "Neto, Filipa"

A mostrar 1 - 1 de 1
  • Profiling airway microbiome composition through volatilomics
    Publication . Neto, Filipa; Ferraz, Ricardo; Vieira, Mónica; Prudêncio, Cristina; Rufo, João; Almeida Vieira, Mónica Andreia; Cavaleiro Rufo, João
    The airway microbiome is known to mediate respi-ratory health. However, current available methods for microbiome analysis are time-consuming and/or rep-resent significant costs for a generalized application in clinical practice. Volatilomics has been suggested as a rapid and low-cost approach to screen microbial pro-files in human samples. Therefore, we aimed to study the efficacy of volatilomics in discriminating microbi-al isolates collected from human breath condensate samples. Bacterial strains showing significant growth under conditions simulating the respiratory environ-ment were isolated. Each strain was standardised to an inoculum of 10⁸ CFU/mL and analysed using an elec-tronic nose equipped with a six-sensor matrix. Data were explored through principal component analysis, cluster analysis, pattern analysis, random forests and recursive partitioning regression. One sensor was re-moved from the analysis due to high collinearity. Prin-cipal component analysis was able to separate strains and the control mainly through the second principal component (p = 0.024), characterized by high MQ3 and MQ8 sensor responses. Sensor profile maps showed distinct volatile patterns across strains (Fig-ure 1), suggesting the presence of distinct microbial signatures. However, reproducibility was low between replicas and time since culture. Recursive partitioning for separating sterile controls from inoculated sam-ples showed the highest accuracy (AUC = 0.73). These results show the potential of separation of microbial strains based on volatilomics. Nonetheless, relative robustness was only achieved for the discrimination of sterile vs inoculated samples.
    2026-05-29Documento de conferência Acesso aberto Ver mais