Advanced method for preparing cell-free DNA sequencing libraries

Period of Performance: 05/01/2017 - 04/30/2018

$270K

Phase 1 SBIR

Recipient Firm

Somagenics, Inc.
Santa Cruz, CA 95060
Principal Investigator

Abstract

The goal of the proposed research is to develop an improved method, RealSeq-DC, to prepare libraries of cell-free DNA (cfDNA) for next-generation sequencing (NGS). cfDNAs, which are found in blood and in other bodily fluids, represent promising, minimally invasive ?liquid biopsy? samples for human cancer and prenatal diagnosis of fetal genetic diseases. cfDNAs comprise highly-fragmented double-stranded DNA fragments, ~50 to 200 bp in length, having single-strand nicks as well as 5'- and 3'-end overhangs. They are normally present at low concentration in biofluids. In cancer patients, concentration of cfDNA and level of fragmentation are positively correlated with tumor weight, with the majority (~80%) of cfDNA fragments being shorter than 100 bp. Analysis of tumor-specific characteristics of cfDNA, such as the amount of DNA, its level of fragmentation, and the presence of mutations and methylated residues can be utilized for cancer diagnosis and prognosis, and for evaluating tumor progression and response to treatment. Next-generation sequencing (NGS) has great potential to assess these parameters. However, due to the high level of DNA fragmentation in cancer-derived cfDNAs, they cannot be efficiently incorporated into DNA-Seq libraries (and, therefore, are under-detected) by conventional, double-stranded methods of sequencing library preparation. To overcome this problem, we propose an advanced method that uses short ssDNA fragments prepared by denaturation of cfDNA for ligation with a single combo adapter followed by circularization of the ligation product and direct PCR amplification (rather than rolling-circle amplification, RCA) of the circular templates. This method produces monomer amplicons each containing a single cfDNA sequence insert flanked by standard Illumina 5'- and 3'-adapter sequences. The method minimizes the formation of amplicons comprising empty ?adapter-dimers?. In Phase I, we plan to develop enzymatic steps specific for ssDNAs, demonstrate the feasibility of the RealSeq-DC approach (proof-of-concept) and its superiority in sequencing DNA fragments of ? 100 nt (the size range which is typical for cancer-specific cfDNA) over the two currently available methods of DNA-Seq library preparation. cfDNA isolated from 4 plasma samples from breast cancer patients will be assayed for this comparison. In Phase II, we will further streamline the RealSeq-DC protocol for commercialization, extend the protocol to identify methylated nucleotides, test its reproducibility and minimize the required cfDNA input. We also will increase the number of samples studied as well as the range of physiological states and diseases with which these samples are associated to evaluate the full potential of this approach and identify any limitations.