October 1, 2021 – Lexington, KY – Today, TruDiagnostic announced their newest research project; expanding and validating a new cellular deconvolution method to separate epithelial cells from immune cells in saliva samples.

Epigenetics is the interface through which our body, at a molecular level, adapts to the world we experience. Epigenetic markers like DNA methylation can react to outside stressors to activate or silence genes, an action which then impacts the rest of the body. 

Cellular deconvolution, also called “cell type composition analysis” or “cell type ratio analysis” is used to estimate the proportions of different cell types collected in a single sample. This is highly useful for epigenetic research, because it allows more accurate examination of the methylation happening in isolated cell types.

Changes to DNA methylation are specific to cell types. Cells specific to the immune system, for instance, have a different methylation pattern than epithelial cells. Cellular Deconvolution lets researchers isolate data from one cell type, instead of pulling combined data from all the different cell types at the same time. 

Additionally, cellular deconvolution can help detect pathological conditions, or predict treatment options. A very low lymphocyte count; a type of cell from the immune system, can indicate a possible infection or other significant illness. 

Saliva is a relatively easy tissue to access. DNA in saliva can be easily extracted, and saliva contains both epithelial cells, and immune cells. 

Validating this method would expand the amount of epigenetic samples available to researchers, because the public is more familiar and comfortable with offering saliva samples for genomic research. Blood as a sample still makes many people uncomfortable to extract or handle, which limits research participation. 

TruDiagnostic and external research partners worked to create a DNA methylation matrix.

A DNA methylation matrix is a way to identify the locations on the DNA where methylation can occur, methylation values, and cell types within a sample. This matrix model, when accurate, could be used as a reference guide for cellular deconvolution. 

 So far, this matrix has proven very accurate in predicting the true cell-type fractions of independent samples when examining 2 cell types. Importantly, the tested performance remained robust while using different technologies, like Stem-Cell-Matrix Compendium-2 (SCM2), 450k technology, BLUEPRINT, and EPIC technology. 

“Almost all epigenetic algorithms have been created via blood cell datasets, which is why blood tissue is widely used for accurate methylation detection and quantification.  However, even in blood, measurements can be highly variable if the different types of immune cell subsets are not controlled for during the analysis. This is where cell deconvolution methods are so vital, as they allow us to control for these variables during analysis.” Says Dr. Dwaraka, Head of Bioinformatics at TruDiagnostic.

Saliva epigenetic samples face the same problem but face more confounding variables due to greater number of epithelial cell types found from the buccal tissue. To alleviate this issue, we have created this saliva deconvolution method to make sure we accurately control for these cell quantities which can change with age and conditions like smoking.  This is a critical step in making sure that the innovations created from blood methylation can also be applied to Saliva tissue samples. 

TruDiagnostic is now working on developing this matrix further, to create and validate a saliva DMRM for 8 cell-types. The 8 cell-type reference would estimate fractions for all 7 immune cell subtypes, in addition to the epithelial cell fraction. This would create a way to closely examine methylation of immune cell types and give more useful clinical data from each saliva sample.