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

Data were analyzed in R. deconvolution analysis by random forest classification was done using the Boruta package, with a maximum of 300 iterations. Protein classifiers that whose mean expression in the test phenotype was significantly greater than alternative phenotypes were taken forward for further analysis. T-SNE analysis was carried out using the Rtsne, with the iterations parameter set to a maximum of 300 and a perplexity value of 30. Analysis was carried out in two dimensional space. Cell type enrichment analysis was done one the enrichR platform. The input search term used for enrichment analysis was the RF protein classifier lists derived from random forest classification. All pairwise comparisons between the expression level of RF classifiers between phenotypes and frequency of immune cell types between survival states was done using t-tests on log10 normalised data. The deconvolution data set that was used to identify phenotype-specific protein classifiers was obtained from a previously published paper by Rieckmann et al. 12