Systems and methods to detect rare mutations and copy number variation

What is this patent about?

’364 is related to the field of molecular diagnostics and bioinformatics, specifically focusing on the high-sensitivity detection of genetic aberrations in cell-free polynucleotides. The technology addresses the inherent challenges of analyzing cell-free DNA (cfDNA), where tumor-derived or fetal-derived genetic material is often present at extremely low concentrations compared to background germline DNA. By integrating advanced sample preparation with computational error-correction, the system aims to distinguish true biological mutations from the noise introduced by PCR amplification and sequencing artifacts.

The underlying idea behind ’364 is the use of a digital sequencing framework that treats individual polynucleotide molecules as discrete information packets to be tracked and verified. The core inventive insight involves non-unique tagging combined with physical molecule characteristics—specifically the start and stop positions of fragments—to create a unique identity for every parent molecule in a sample. This allows the system to collapse multiple sequencing reads into a single high-fidelity consensus sequence, effectively filtering out stochastic errors and providing a precise count of original molecules to detect both rare mutations and structural variations.

The claims of ’364 focus on a specialized system for detecting somatic genetic variants in cfDNA from human blood samples using a combination of molecular barcoding and fragment geometry. The system utilizes a nucleic acid sequencer to process cfDNA molecules that have been joined at both ends with molecular barcodes from a predefined set, where the number of available barcodes is intentionally fewer than the number of molecules mapping to a specific genomic position. The claimed processor then identifies unique parent molecules by cross-referencing these barcode sequences with the specific beginning and end base positions where the reads align to a human reference genome.

In practice, the invention functions by grouping sequencing reads into familial sets derived from the same original parent molecule. The system specifically filters for fragments with a length of 140 to 180 nucleotides, a characteristic size range for cfDNA, to ensure the analysis targets the most relevant biological material. By comparing the sequences within these familial groups, the system can detect a broad spectrum of somatic variants, including single nucleotide variants (SNVs), copy number variations (CNVs), indels, and gene fusions, even when these variants occur at frequencies below the nominal error rate of the sequencing platform.

This approach differs from prior methods by maximizing conversion efficiency and utilizing a hybrid identification strategy that does not rely solely on unique barcodes. Traditional sequencing often loses a significant percentage of the starting material during library preparation, but this system is designed to capture and sequence the vast majority of molecules in a 10 mL blood draw. By leveraging the natural diversity of fragment breakpoints alongside a limited set of barcodes, the technology achieves the sensitivity required for early-stage cancer monitoring and treatment adjustment without the prohibitive complexity of billions of unique tags.