Storing biological samples is a risky affair. A single error in the handling, processing or labeling step can give inaccurate data or render biosamples unusable for research. Because of these high stakes, Pellerin et al. developed a genotyping method to detect errors in handling by validating the sample identity.
First, the team selected seven variable number tandem repeats (VNTRs) to analyze: D16S83, D17S5, D1S80, D19S20, D1S111, D17S24 and RB1. For each locus, they optimized polymerase chain reaction (PCR) conditions using unrelated human genomic DNA, to allow clear detection of the PCR products by agarose gel and ethidium bromide staining. Next, they developed a DNA-profiling algorithm using DNA extracted from the buffy coat fraction of 23 unrelated individuals. The profiling algorithm consisted of a combined profile of the three most discriminant loci: D16S83, D17S5 and D1S80. Taking this combined profile into account, these 23 individuals had a match probability of one in 1,637. Then, following the first profiling step and turning to the indistinguishable individuals, the team continued with the next steps of the algorithm, this time using the four other VNTR loci. The researchers evaluated each locus one at a time until they could distinguish between individual samples. As each step progressed, the calculated match probability decreased significantly, and employing all seven loci resulted in a match probability of one in more than 7 million.
Putting their method to the test, the team surveyed an additional 101 randomly selected cases, totaling 403 DNA samples, from the PROCURE Biobank cohort. This biobank includes tissue, blood and urine samples from prostatectomy patients receiving care at four Quebec university health centers. The patients provided prostate tissues at the time of surgery and blood at three different time points over a period of up to seven years, with the exception of one patient. The researchers extracted DNA from blood and/or prostate tissue. They used a NanoDrop ND-1000 spectrophotometer (Thermo Scientific) to determine the concentration of the DNA samples.
In the 403 samples, the team identified two mismatches. In one case, they determined that two out of three DNA samples from a single patient showed differences, and the two samples also had an entirely different DNA profile compared with others in the study. After further investigation, they found that the mismatch occurred when the wrong block of tissue was chosen from the −80°C freezer. Profiling of the DNA extracted from the correct tissue confirmed that the D16S83/D17S5/D1S80 profile matched the patient’s blood sample.
The second error occurred from mislabeling. Biobank worksheets reversed labels from tissue blocks belonging to two patients who underwent surgery on the same day. In this situation, DNA profiling confirmed the error between these two cases.
The authors note that the DNA profiling resulted in a 0.5% error rate, which was still within accepted limits. Still, the discovery of these errors validates the value of using genotyping as quality assurance in biobanking.
1. Pellerin, C., et al. (2016) “A simple variable number of tandem repeat-based genotyping strategy for the detection of handling errors and validation of sample identity in biobanks,” Biopreservation and Biobanking, 14(5) (pp. 383–389), doi: 10.1089/bio.2015.0113.