Subtractive metabolomics between microbial secretions with differential CBI activity.

<p>The metabolite profiles of <i>B. sp.-A</i> (blue), <i>B. sp.-B</i> (green), and two other CBI-inactive microbial secretions (red and aqua) were used to generate a PCA plot of secreted metabolite composition (A). Clustering of the replicates supported the separation of the two <i>B. sp</i>. strains and highlights the potential existence of one or more biomarkers that facilitate the characterization of the bioactive secretion components. An S-plot (B) was then used as a multivariate analytical technique to identify putative biomarkers by combining covariance and correlation measures for all analyses together in a scatter plot. Variables included in the matrix that separated compounds were exact mass, abundance and retention time. The y-axis indicates the confidence, or reliability of a measure and the x-axis reflects the magnitude of the variable. Note that features falling on the “tails” of the S-plots are most likely to be biomarkers that are differentially present in samples. Differentially abundant compounds which represent potential CBI compound candidates are highlighted in red boxes.