BACKGROUND: Unsupervised clustering of biomarkers derived from non-invasive samples such as nasal fluid is less evaluated as a tool for describing asthma endotypes. OBJECTIVE: To evauate whether protein expression in nasal fluid would identify distinct clusters of asthmatics with specific lower airway molecular phenotypes. METHODS: Unsupervised clustering of 168 nasal inflammatory and immune proteins and Shapley values was used to stratify 43 severe asthmatic patients (ENDANA) using a two 'modelling blocks' machine learning (ML) approach. This algorithm was also applied to nasal brushings transcriptomics from U-BIOPRED. Feature reduction and functional gene analysis were used to compare proteomic and transcriptomic clusters. Gene set variation analysis (GSVA) provided enrichment scores (ESs) of the ENDANA protein signature within U-BIOPRED sputum and blood. RESULTS: The nasal protein ML model identified two severe asthma endotypes, which were replicated in U-BIOPRED nasal transcriptomics. Cluster 1 patients had significant airway obstruction, small airways disease, air trapping, decreased diffusing capacity and increased oxidative stress, although only 4/18 were current smokers. Shapley identified 20 cluster-defining proteins. Forty-one proteins were significantly higher in Cluster 1. Pathways associated with proteomic and transcriptomic clusters were linked to Th1, Th2, neutrophil, JAK-STAT, TLR and infection activation. GSVA analysis of the nasal protein and gene signatures were enriched in subjects with sputum neutrophilic/mixed granulocytic asthma and in subjects with a molecular phenotype found in sputum neutrophil-high subjects. CONCLUSIONS: Protein or gene analysis may indicate molecular phenotypes within the asthmatic lower airway and provide a simple, non-invasive test for non-T2 asthma that is currently unavailable.