De-convolution of Body Fluid Mixtures: Cell Type Identification and Genetic Profiling of Micro-Dissected Cells.

摘要

Mixtures are frequently encountered in forensic DNA analysis due to the involvement of multiple persons during the commission of certain crimes. Mixtures are also often found associated with ‘touch DNA’ or with other crime scene samples due to unavoidable contamination with previously deposited DNA. Analysis and interpretation of these admixed DNA samples is often quite complex and therefore extremely challenging even more so if the number of donors exceeds two. In certain situations, the presence of a mixture can be anticipated and the cells separated prior to analysis, such as is the case with a ‘differential extraction’ of sexual assault evidence where sperm cells are present. In many cases, however, the cellular components of a mixture cannot be separated due to the presence of non-easily distinguishable epithelial cells or leukocytes from more than one donor. Interpretation of these mixtures is sometimes further complicated because of degraded DNA or artifacts that arise during the PCR-based analyses. A limited number of tools are available to operational forensic laboratories in order to aid in the interpretation or handling of mixtures. The current work sought to provide a novel strategy for the de-convolution of body fluid mixtures containing non-distinguishable cell types (e.g. epithelial cells and leukocytes) through an isolation of individual cells using laser capture micro-dissection and recovery of genetic material (RNA and DNA) for cell type identification (RNA profiling methods) and donor identification (autosomal STR analysis). Thus the cells comprising mixtures would be separated out according to body fluid source and single source profiles obtained from each of the individual mixture contributors. Uniquely, such an approach would permit the attribution of each DNA profile in a mixture to a particular contributing body fluid. In this study, we demonstrate the ability to isolate single epithelial cells using laser capture micro-dissection (LCM) and also using a novel micro-manipulation approach. We have developed highly robust and sensitive DNA lysis and amplification strategies, including the use of micro-volume PCR reactions as well as single-tube combined lysis and amplification reactions, for the enhanced analysis of single (or few) epithelial cells isolated from admixed epithelial cell mixtures which can aid in the de-convolution of epithelial cell admixtures. While the routine use of single cell profiling still requires additional work, we also report the development of a novel strategy for a more efficient and informative genotyping of two person body fluid mixtures containing non-distinguishable cell types (e.g. epithelial cells and leukocytes) at approximately equal concentrations through a physical isolation of multiple samplings of groups of cells using LCM and subsequent DNA profiling of the cell population samples. Genotype inference and the attachment of statistical weight are accomplished via the TrueAllele? quantitative computer interpretation system. The combination of multiple binomial sampling of the total mixed cell population to create separate sub-populations with differing weight ratios, together with quantitative computer-based statistical modeling, can provide a significant information gain compared to the standard analytical approach. Additionally, we also report the development of preliminary RNA profiling strategies to provide a determination of the cell type of origin of epithelial cells (buccal, vaginal and skin), which permits further individualization of epithelial cells from admixed samples.