001, * p < 0.01, ns: not significant. Genotyping of host microsatellite markers To measure host population structure we amplified ten polymorphic microsatellite loci [29] covering nine linkage groups. These included Cg157 (LG I 141.5 cM), Cg_194 (LG I 28.7 cM), Cg136 (LG II), Cg193 Everolimus (LG III), Cg_164 (LG IV), Cg173 (LG V), Cg_i28 (LGVI), Cgi24 (LG VII), Cg172 (LG IX) and Cg_181 (LG X) [30]. Loci were amplified using M13-tail labelling [31] in 20 μl volume using 4 μl of 5x concentrated buffer containing MgCl2 (Promega,

Mannheim), 10 nM of each dNTP, 13.9 μl of H2O, 5 nM of each locus specific primer, 8 pmol of one M13-tail labelled with either FAM, VIC, NED or PET fluorophores and 1 unit of GoTaq Polymerase (Promega, Mannheim). Cycling used a standard protocol consisting of 5 min

at 94°C followed by 28 cycles at 94°C (30 s) / 56°C (45 s) / 72°C (45 s). M13 tail incorporation was achieved in 8 cycles at 94°C (30 s) / 53°C (45 s) / 72°C (45 s), and a final extension at 72°C for 10 min. PCR products were pooled into sets of four loci with differently coloured labels and separated on a ABI Prism 3130 XL (LifeTechnologies, Darmstadt) capillary sequencer using a LIZ 500 size standard. Product sizes were scored using the GeneMarker v1.90 software (SoftGenetics) and pairwise genetic differentiation between populations was calculated as Wright’s fixation indices (FST) according to Weir & Cockerham [32]. Pacific oysters are known to harbour substantial amounts of null alleles [33] that could bias any estimate of population differentiation. C646 mw We therefore estimated the frequency of null alleles within our sample using the software freeNA [34]. Frequencies of null alleles were small for all loci and populations (range: 0 – 0.06) except for locus Cgi-194 where estimates were higher within Levetiracetam all oyster beds (range: 0.05 – 0.15). We therefore excluded this locus from the analysis, and only report the corrected F ST values after removal of loci with high frequencies of null alleles in all populations. Genetic distance between

individuals was calculated as geometric AMOVA distances: [35], where distances between individual genotypes j,k are summed over S loci. Calculations were performed using the R package Gstudio. Amplicon sequencing of microbial communities Microbial diversity and composition was measured within a standardised amount of genomic DNA (30 ng). We used an informative part of the 16S rRNA gene spanning position 535-904/912 in the E. coli genome covering the variable regions V3 and V4 for ribotyping. Using a PCR product of this length increases the precision of taxonomic assignment [36] and should provide high resolution due to the inclusion of two variable regions. Initial testing of these primers revealed that they preferentially amplified host 18S rRNA (24 out of 24 randomly picked clones).