Comparability of genotyping data obtained by different procedures an inter-laboratory survey

Characterisation of germplasm resources by applying molecular markers is increasingly becoming general practice, but there are few studies on the feasibility of sharing and exchanging genotyping data generated by different laboratories. The present study was aimed at comparing PCR-based simple sequence repeat (SSR) or microsatellite profiles of apple cultivars, obtained independently by two research groups. Each group analysed an identical set of 61 DNA samples at the same five microsatellite loci, using their own laboratory procedures. A comparison of the two datasets revealed that 99.3% of the single-locus genotypes corresponded at one of the alleles, at least. However, three types of discordance were observed: (i) 'dropout of longer alleles'; (ii) mis-scoring of stutter patterns as homozygous or heterozygous; and, (iii) complete allele mismatch. These inconsistencies occurred with frequencies of 14.8%, 2.0% and 0.7% across all single-locus genotypes. It is concluded that the rate of genotyping errors could be reduced by an a-priori standardisation of PCR protocols across different laboratories. In order to provide inter-study conformity, and comparability of data, we further propose that allele sizes be displayed as absolute values, after having applied a simple conversion procedure based on the exact allele sizes in standard cultivars.