Ultrasensitive nucleic acids analysis using digital sequencing

Abstract

Liquid biopsies offer a minimally invasive and affordable procedure to obtain valuable clinical insights through biomarker analysis that can aid to manage the care of different diseases and health conditions. Analysis of cell-free DNA that is shedded into blood plasma from different tissues and cell types is of particular interest, providing molecular information that can be used in clinical decisions. In the case of cancer, cell-free DNA includes a fraction of DNA liberated by tumoral cells, containing tumor-specific mutations. Analysis of circulating tumor DNA has applications like monitoring disease progression and determining therapy effectiveness to avoid under- or over treatment. However, in most clinical applications the levels of circulating tumor DNA are very low, hence it cannot be assessed with traditional analytical methods. In this thesis, we aimed to develop and optimize the SiMSen-Seq workflow, which is a digital sequencing method that enable ultrasensitive detection of DNA variants using molecular barcodes, also enabling variants detection in RNA. First, we studied the effects of preanalytical variables, including the type of blood collection tube, time between sampling and plasma isolation using blood collected from healthy individuals. Our results indicate that most preanalytical parameters are essential to consider in order to obtain high cell-free DNA yield and avoid cellular contamination. Then, we designed and tested the effects of structured elements in the molecular barcode that is used bioinformatically to correct errors in the sequencing data. Our results proved that assays using structured molecular barcodes outperformed those using standard non-structured barcode sequences, resulting in improved sensitivity and reduced amounts of off-target reads. Finally, we adapted the SiMSen-Seq workflow to include a reverse transcription step that enables the detection of RNA target sequences. Our data demonstrate the importance of using a reverse transcriptase and reaction condition that generate high yield of complementary DNA and produce few enzymatic sequence errors. This approach can detect RNA variants, including modifications, in clinical samples with the same sensitivity as for DNA. In conclusion, our protocols and data contribute to the field of liquid biopsy analysis and the ability to detect DNA and RNA molecules as biomarkers. We expect that biomarkers analysis in blood will be increasingly important in clinical decision and disease management with the goal of improving treatment outcome and quality of life.