Fig. 3
Dynamics of nasal microbiota profile membership. The number of samples in each cluster at each time point was visualized in alluvial diagrams, which were stratified by pneumococcal carriage3 outcome. Cluster membership was determined using average linkage hierarchical clustering based on the Bray–Curtis dissimilarity matrix. Clusters were characterized by Staphylococcus (2; STA); Corynebacterium (3) and Dolosigranulum (4; CDG); Corynebacterium (1; COR); Haemophilus (9; HPH), Moraxella (6; MOR), Fusobacterium (10; FUS), Streptococcus (7; STR), and Peptoniphilus (5), Finegoldia (8), Anaerococcus (11) and Streptococcus salivarius (13; PEP/MIX). The dynamics of nasal microbiota profile membership were shown for high-dense carriers (a), low-dense carriers (b) and non-carriers (c). The height of the figures corresponds with the total number of samples within that group. In addition, the height of the nodes and the thickness of the edges connecting the nodes is proportional to the number of samples. The number of samples in each cluster at each time point, stratified by carriage3 outcome, is provided in Supplementary Table 5. The number of cluster changes was lower in low-dense carriers compared to both high-dense and non-carriers (see Tables 2 and 3). A higher proportion of CDG profiles was observed in low-dense vs. non-carriers. Contrariwise, low-dense carriers specifically lacked STA-dominated profiles compared to both high-dense and non-carriers
