Skip to main content
Fig. 2 | BMC Oral Health

Fig. 2

From: Exploring the microbiome of oral epithelial dysplasia as a predictor of malignant progression

Fig. 2

Prevalence, abundance and differential abundance of the top 25 most abundant genera in progressor and non-progressor samples. Phylogenetic tree showing the class of each genus and a heatmap showing mean prevalence (blue scale; left). Heatmaps showing mean abundance of genera in progressor (P) or non-progressor (NP) samples (left) or differential abundance (centre) are shown alongside boxplots showing abundance in all samples (right) for each of relative abundance and CLR abundance. In the boxplots, each sample is shown as an individual point and boxes show the median, upper and lower quartiles while whiskers show the range of the data (1.5 times the interquartile range). MaAsLin2 tests for differential abundance were run with (+ M) and without the other clinical variables (metadata, M; age, sex, ethnicity, alcohol intake, smoking status, grade of dysplasia and lesion site). As above for PERMANOVA tests, progression was grouped in one of three ways: (i) Ps and NPs (P vs NP); (ii) Ps grouped by the number of years to progression (< 1, 1–2, 2–3, 3–4, 4–6 and 6 + ; P vs NP); and (iii) Ps only with the number of months to progression (P only). For (i) and (ii) the matched P/NP grouping was given to the model so only matched controls were used. Genera were determined to be differentially abundant and are shown in black in the heatmap if they had q ≤ 0.25 (the default in MaAsLin2). White denotes that they were not significantly differentially abundant

Back to article page