Accelerating the Development and Validation of Liquid Biopsy via Computational Error Suppression

The ability to identify low-frequency genetic variants among heterogeneous populations of cells or DNA molecules plays a pivotal role in many oncology applications such as minimal residual disease detection and liquid biopsy. Central to this endeavor is to enhance the accuracy of next generation sequencing (NGS) technologies. Current studies have largely focused on experimental error suppression approaches such as barcode-based consensus sequencing methods. By contrast, computational error suppression approaches are lacking. We recently developed methods that suppress NGS error rates to 100-fold lower than published reports. Based on our extensive experiences in analyzing the genomes of >2,000 childhood cancer patients, we propose to develop novel computational approaches for sensitive analysis of low-frequency variants. We will apply these approaches to i) studying the clonal diversity of the diagnosis sample, ii) track the residual tumor in remission samples and iii) monitor the emergence of drug-resistant variants during treatment in leukemia patients. Our research outcome will significantly enhance the sensitivity of detecting low-frequency variants and will also lead to objective frameworks for benchmarking detection methods in clinically-relevant manner. These innovations will accelerate the development and validation of liquid biopsy for cancers.

Read more about this work here: Applications of probability and statistics in cancer genomics