3.3 Candidate gene analysis
The 96 candidate immune genes (out of 166) that remained in our second HDMKPRF analysis include 13/17 recognition genes, 21/35 signalling genes, 4/21 effectors, 16/35 protease coding genes, 24/31 haematopoiesis genes, 13/18 melanisation genes and 5/9 wound healing genes (Table S2). Seven of these immune genes were found to be under positive selection inG. pusilla , which was significantly more than the single gene identified under positive selection in G. calmariensis (Fisher’s Exact test: Odds ratio = 0.10, p<0.02), but not significantly greater than the four genes under positive selection in G. tenella (Odds ratio = 0.61, p>0.3), when related to all genes under positive selection in respective species. However, while the number of genes under positive selection in G. tenella was marginally higher than the genes under positive selection in G. calmariensis , this was not significant (Odds ratio = 0.17, p<0.09). Positively selected immune genes in G. pusillainclude genes involved in parasitoid recognition (santa-maria andCorin ), Toll and JNK pathways (grass and Tak1 ), a protease with serine-type carboxypeptidase activity and genes involved in lamellocyte differentiation (cher and zfh1 ) (Table 2, Table S6). The four immune genes under positive selection in G. tenella included two recognition genes (Corin andPGRP-LE ) and two genes with important roles in haematopoiesis through regulation of lamellocyte differentiation (cher andCyt-b5 ) (Table 2, Table S6). Finally, the single gene under positive selection in G. calmariensis was Cyp9f2 , which is involved in melanization.