Skip to main content
Fig. 1 | BMC Biology

Fig. 1

From: Physiological constraints dictate toxin spatial heterogeneity in snake venom glands

Fig. 1

The venom gland of N. nigricollis contains functionally distinct clusters of related toxins. A Detected toxins can be grouped into three clusters with distinct distributions across the venom gland of N. nigricollis. The top spectrum shows the normalised across-tissue averaged spectrum in the m/z region corresponding to 3FTx—the dominant components of N. nigricollis venom (full spectrum in Additional file 1: Fig. S1). The below centroid spectra are extracted from each group clustered by probabilistic latent semantic analysis (pLSA), where distributions across the gland are displayed as contrast-optimised heatmaps on the right. The top image shows the section stained with haematoxylin and eosin (H&E) post-MSI acquisition, and the direction of venom secretion is indicated by the arrow. B Toxin distributions are correlated with functional and phylogenetic relationships, as shown by comparing the phylogenetic relationship and distributions of identified 3FTx. The displayed 3FTx phylogeny was reconstructed by maximum likelihood (ML) under the FLU+G4 model (left, bootstrap support at nodes, displayed as mid-point rooted), while their distribution across the venom gland as determined by MSI is shown as contrast-optimised heatmaps on the right. C The relationship between function, relatedness, and distribution is further supported by a significant correlation between pairwise ML distances and spatial correlations of the identified 3FTx (left; Spearman rank: P < 0.001, r = − 0.337, 95% CI = − 0.491 to − 0.163, R2 = 0.1944) and significantly higher spatial correlations among toxins within compared to between each functional class (right; Mann-Whitney two-tailed: P = 0.0481 and < 0.0001 for neurotoxins and cytotoxins, respectively). Heatmap legend is shown in A, and scale bar represents 2 mm

Back to article page