Proteome fingerprints define groups with distinct clinico-pathological phenotypes in the U-BIOPRED asthma study
European Respiratory Journal, 2016
Background: Severe asthma is not a single disease and likely consists of several subtypes (phenot... more Background: Severe asthma is not a single disease and likely consists of several subtypes (phenotypes) but the molecular basis for this diversity is poorly understood. Aims: The pan-European U-BIOPRED consortium utilized state of the art omics technologies to characterise the molecular biology of severe asthma sub-phenotypes as 9omic fingerprints and combined 9omic handprints by systems biology. Methods: 250 induced sputum samples from 169 severe asthmatics (49 smokers), 40 mild to moderate asthmatics and 41 healthy control participants were analysed by label-free quantitative data independent mass spectrometry. Abundance profiles of proteins between participants was analysed by topological data analysis (TDA) supported by standard statistics. Topological graphs of the proteomic data allowed us to identify groups of participants defined by their similarity in sputum proteome profiles (fingerprints); their clinico-pathological characteristics were defined and comparisons were made between groups. Results: In asthmatic participants, 10 groups were identified with distinct proteome fingerprints (proteomic phenotypes) and similarly diverse clinico-pathological phenotypes and were associated with different granulocyte counts: 4 neutrophilic, 3 eosinophilic, 2 paucigranulocytic and 1 mixed granulocytic. Machine learning further identified biomarkers predictive of each of the 10 fingerprints. Conclusions: Characterisation of the proteome fingerprints of asthma phenotypes, and the identification of phenotype-specific biomarkers represents a step towards stratified medicine and personalised treatment of asthmatics.
Uploads
Papers by Dominic Burg