Null alleles of microsatellites for Manila clam Ruditapes philippinarum

Source/description: The Manila clam Ruditapes philippinarum (synonym Venerupis philippinarum1) originates from the Indo-Pacific region and has been introduced worldwide for aquaculture and fisheries.2 As in many other species, microsatellites seemed to be promising markers for population genetics. For the Manila clam, the presence of microsatellite null alleles was reported, but this has not been checked by specific tests. Sample: :Five introduced R. philippinarum populations (Figure S1) from Italy, Spain and Portugal were genotyped in order to assess the occurrence of null alleles and to test markers reliability. Microsatellites analyses: Samples were genotyped for seven microsatellites: Asari16, Asari23, Asari24, Asari543 and Ktp5, Ktp8, Ktp224 as previously described.5 Summary statistics calculated using GENALEX 6.56 showed that all loci were polymorphic in each population (Table S1). However, all loci were affected by failed amplifications. For Asari16, 5% of samples did not amplify, but percentages were higher at other loci: 25%, Asari23; 27%, Asari24; 60%, Asari54; 54%, Ktp5; 35%, Ktp8; and 45%, Ktp22. No correlation with DNA quality, PCR procedures and missing amplifications was observed, suggesting the occurrence of homozygous null allele genotypes. Thus, the seven loci were scored for null allele occurrence7,8 by means of the programs MICRO- CHECKER 9 and FREENA 10 (http://www.montpellier.inra.fr/ URLB/). MICRO-CHECKER incorporates four different algorithms (Oosterhout, Chakraborty, Brookfield 1 and Brookfield 2) and is recommended as a standard procedure.11 The results suggest that all loci, except Asari16, are affected by null alleles in all populations (Table S1). In 49% of the Marano Lagoon samples, missing amplifications occurred, which precluded analysis by MICRO-CHECKER but not by FREENA (Table S1), and interfered with the correction of allele frequencies.12 We used the ENA (excluding null alleles) method in FREENA to correct allele frequencies. The Wilcoxon test showed highly significant differences between corrected and uncorrected allele frequencies (P < 0.001) and between FREENA-corrected and uncorrected observed heterozygosity (P < 0.001). Furthermore, for most loci, we observed a negative correlation between the frequency of null alleles and allelic diversity (Figure S2). However, neighbor-joining trees using the PHYLIP package on the basis of uncorrected and FREENA-corrected data were similar (Figure S3a, b). These findings agree with the previous suggestions13 that allele frequency corrections allow the evaluation of population differentiation levels (FST) but do not lead to a reliable assignment of individuals to populations.