When it comes to brushing, whether it’s your teeth or a cheek swab sample, proper technique is important. Your dentist may have already told you not to brush you teeth too hard, but it turns out how you collect a cheek swab is also important. Buccal epithelial cells are perhaps the most clinically accessible tissue source. As such, they have been used frequently for genetic and epigenetic studies. While the source tissue for genetic studies is usually not consequential, epigenetic changes are highly cell-type specific. Numerous studies of DNA methylation from buccal and saliva have shown a relatively large amount of variation compared to other tissues.
The laboratories of Michael Brudno and Rosanna Weksberg at the Hospital for Sick Children (Toronto, Canada) sought to identify and correct for blood cell contamination in buccal/saliva sample DNA methylation data. Using six blood cell types and purified buccal cell DNA methylation data as a reference, they estimated cell-type composition for 11 published DNA methylation studies that used the 450k array. They estimated the buccal purity and blood cell contamination for each study and looked for possible explanations for low-purity studies.
Here’s what happened:
- They developed a reference-based method to estimate and correct for buccal cell data heterogeneity that removes variance due to data impurity while preserving cross-tissue differences
- This method generates a new corrected set of DNA methylation values, similar to batch effect correction
- Applying the correction led to much better separation of buccal samples from blood samples while preserving variation within the buccal sample group
- Most of the datasets contained a subset of low-quality (blood cell contaminated) buccal samples
- There were no low purity samples in a study by Martino et al. which used a different swab type (Catch-All Sample Collection Swabs) with softer spongelike brushes which may be associated with less bleeding
- Several studies had significantly decreased buccal data quality in samples from patients with neurodevelopmental disorders (Downs syndrome, ASD, and FASD) possibly due to more challenging collection of buccal tissue from these patients
First author Andrei Turinsky shares: “Cheek swabs seem easy to collect, which is why they are used widely in genetics and epigenetics. But if you look closely, this apparent simplicity is deceptive. In this study we wanted to lift the curtain a little bit and show that lower-purity buccal samples are too common to be ignored.” Given how common low-quality buccal samples appear to be, and the large amount of variability within them, correction methods such as this one may become a standard processing step in DNA methylation studies moving forward.
Check out the full cheeky story in Epigenetics, March 2019