In-silico immune cell deconvolution of the airway proteomes of infants with pneumonia reveals a link between reduced airway eosinophils and an increased risk of mortality

Charles J Sande; Jacqueline M Waeni; James M Njunge; Martin N Mutunga; Elijah Gicheru; Nelson K Kibinge; Agnes Gwela

The proteomics data reported in this paper are available at the ProteomeXchange Consortium database (Accession numbers: PXD009403) . In silico deconvolution analysis was based on high resolution mass spectrometry data obtained from the airway. Cells were obtained from the naso and oropharyngeal sites using separate swabs, which were both eluted in a common transport media. As shown in A above, these samples contained a broad diversity of immune cells including T-cells, monocytes/macrophages and neutrophils. These cells were isolated from the sample by centrifugation and processed for mass spectrometry analysis using a standard protocol (see methods). (C) In silico phenotype deconvolution was conducted initially by identifying phenotype specific protein features in a previously-published deconvolution data set using random forest feature 24 selection. These features were then applied on airway proteome data set obtained from children admitted to hospital with severe pneumonia, and used to construct a detailed map of airway immunophenotypes and their associations with clinical outcomes. to visualize the separation of immune cell phenotypes at the major phenotype level based on protein expression data from the deconvolution data set. We observed that expression profile of certain cell types including neutrophils, monocytes and basophils, resulted in a clear separation from other immune cell types. To identify specific proteins that could be used to distinguish phenotypes, we used random forest feature selection. Using this approach, we identified protein features for 33 phenotypes that could be used to disaggregate individual phenotypes from a complex mixture (Supplementary table 1 FTH1 GABARAP GNAQ GUSB HLTF IL1B IQGAP3 ITGA5 LACC1 LAMTOR4 LTB4R LYZ METTL7B MPP1 MTFMT NCSTN NINJ1 OBFC1 PAM PLAC8 RAD9A RALGAPA1 RBM47 RIT1 RPL6 SERPINB1 SERPINB13 SERPINB2 SLC44A1 SLC7A6 STEAP3 SUN1 TMEM50A VCAN ZNF185 Airway expression level (log2 reporter intensity) 4