Methods and systems for detecting genetic variants

What is this patent about?

’858 is related to the field of molecular biology and genetic diagnostics, specifically focusing on the high-sensitivity detection of genetic variants in cell-free DNA (cfDNA). The technology addresses the challenges of analyzing heterogeneous genomic samples, such as blood plasma containing circulating tumor DNA, where rare mutations or copy number variations must be distinguished from sequencing noise and amplification artifacts.

The underlying idea behind ’858 is that the accuracy of genetic quantification can be significantly improved by tracking the recovery of both strands of a double-stranded DNA molecule. By using duplex tagging to uniquely identify the original Watson and Crick strands, the system can distinguish between true biological variants present on both strands and errors introduced during the lab process. Furthermore, the invention utilizes the ratio of detected strand pairs to single strands to mathematically infer the number of molecules that were present in the original sample but lost during sequencing.

The claims of ’858 focus on a method for analyzing cfDNA by ligating library adaptors to double-stranded fragments using a high molar excess—specifically more than 10×—to ensure high conversion efficiency. The process involves amplifying these tagged molecules and using the molecular barcodes to sort the resulting sequence reads into families. These families are then categorized based on whether they represent both strands of the original duplex or only a single detected strand, allowing for a more precise reconstruction of the starting sample composition.

In practice, the invention achieves high efficiency by ensuring that at least 20% of the cfDNA population is tagged at both ends. Once sequenced, a programmed computer processor maps the reads to a reference genome and groups them into families using a combination of the barcode sequence and the genomic start/stop positions of the fragments. This multi-layered identification allows the system to collapse redundant reads into a single consensus sequence, effectively filtering out the 'stray' errors that typically plague deep sequencing assays.

This approach differs from prior methods by moving beyond simple redundancy reduction to a more sophisticated statistical estimation of unseen molecules. While traditional techniques only count the molecules they successfully sequence, ’858 accounts for the highly variable recovery rates across different genomic regions. By incorporating these 'unseen' counts into the final analysis, the method provides a much more accurate measure of copy number variation and achieves the extreme specificity required to detect rare cancer-associated mutations at frequencies below 1%.