Systems and methods to detect rare mutations and copy number variation

What is this patent about?

’796 is related to the field of genetic analysis and medical diagnostics, specifically the detection of rare mutations and copy number variations (CNVs) using cell-free DNA (cfDNA). The technology addresses the challenge of identifying low-frequency genetic signals—such as those shed by tumors or a fetus—within a vast background of normal germline DNA found in bodily fluids like blood or plasma.

The underlying idea behind ’796 is to maximize the conversion of scarce starting material into a sequence-ready library while minimizing the noise introduced by PCR and sequencing errors. By achieving a high conversion efficiency during the initial tagging phase, the system ensures that rare molecules are not lost before they can be amplified. This allows the system to distinguish true biological variants from technical artifacts by tracking progeny reads back to their original parent molecules.

The claims of ’796 focus on a specific preparation method that ensures at least 20% of the initial cfDNA molecules are successfully ligated to adaptors at both ends before any amplification occurs. This is achieved by using a significant molar excess of adaptors (more than 60X) relative to the DNA sample. The process involves generating adaptor-linked parent polynucleotides, amplifying them to create progeny, and then selectively enriching or amplifying specific regions of interest for targeted sequencing.

In practice, the invention works by taking a limited sample of cfDNA (often less than 100 ng) and utilizing high-efficiency ligation to create a diverse library of tagged parent molecules. Following a universal amplification step, the system employs selective enrichment to focus sequencing depth on actionable oncogenes or other genomic regions of interest. This targeted approach allows for the detection of mutations at frequencies as low as 0.1%, which is typically below the raw error rate of standard sequencing platforms.

This methodology differs from prior approaches by prioritizing the preservation of the original molecule diversity through extreme adaptor excess and high-yield ligation chemistry. Traditional methods often suffer from low conversion rates (1-5%), which effectively discards the rare tumor-derived fragments needed for early-stage detection. By combining high conversion with consensus sequencing—where multiple reads of progeny are collapsed to verify the parent sequence—the invention provides a high-fidelity digital representation of the patient's genetic landscape.