Skip to main content
Fig. 5 | Genome Biology

Fig. 5

From: The genome of the yellow potato cyst nematode, Globodera rostochiensis, reveals insights into the basis of parasitism and virulence

Fig. 5

Effectors in G. rostochiensis are grouped into ‘islands’. a Dorsal (black) or subventral (grey) effectors are skewed towards a higher neighbouring gene distance compared to random (Student’s t-test, p <0.01), yet are contained within gene dense regions of the genome. b The presence of effectors in adjacent (n ± 1), or neighbouring positions (up to ±9), was determined. As a negative control, a subset of 612 G. rostochiensis gene families not predicted to contain effectors was identified from the OrthoMCL. Starting from this initial negative set of 612 gene families, 37 of these gene families were selected at random and the presence of genes from these 37 families in adjacent (n ± 1), or neighbouring positions (up to ±9), was determined. This process was repeated for 1000 iterations to generate a robust negative for the average frequency in each neighbouring position. The observed frequency of effector occurrence at each position (black bars) was compared to the average of 1000 iterations for the negative (white bars). For non-effector containing gene families, increasing distance from each gene reduces the likelihood of identifying another member in any of the same families (error bars indicate standard deviation of 1000 iterations). The clustering of effector loci extends beyond immediate neighbours, with an excess of effector loci as next-but-two neighbours (n ± 3) and also at n ± 6 (χ2, p 0.01 and 0.001, respectively). c Example of one island (7) at the edge of scaffold 00141. With the exception of high effector density (red), no obvious genetic features are associated (gaps (Ns, purple), AT content (black line), gene density (blue) microsatellites (orange), variants (black bars) and transposable elements (TEs, green)). d No difference in transposon density was found between islands, in the remainder of scaffolds containing islands, in entire scaffolds containing islands or in scaffolds numerically adjacent to those containing islands (Kruskal–Wallis, p = 0.515, error bards indicate standard deviation). When each island is treated as a single locus, the nearest external transposable element 5′ of the first gene, and 3′ of the last, is significantly closer than expected (ANOVA, n = 39, p = 0.028 accounting for multiple testing, Fig. 5d). Interestingly, the inverse measurement (the closest internal transposon to each island border), is not significantly closer than expected (n = 45, p = 0.116, Fig. 5d), suggesting that this may be a feature of islands as an integral whole, rather than the separate genes comprising the islands

Back to article page