Obviousness

The obviousness of US patent 11149306 under 35 U.S.C. § 103 can be analyzed by considering combinations of prior art references that would motivate a person having ordinary skill in the art (PHOSITA) to arrive at the claimed invention. The priority date for US11149306 is December 28, 2013.

The "Background" and "Summary" sections of US11149306 highlight the challenges in accurately detecting and quantifying rare genetic alterations, especially in heterogeneous genomic samples like cell-free DNA (cfDNA). The patent states that existing methods, while reducing errors in converted and sequenced molecules, "are not able to infer the counts of molecules that were converted but not sequenced." The claimed invention addresses this by tagging both complementary strands of individual double-stranded DNA molecules, differentiating between "Pairs" (both Watson and Crick sides recovered) and "Singlets" (only one half recovered), and using these counts to estimate the number of unseen molecules, thereby improving sensitivity and specificity.

The patent explicitly incorporates by reference several prior art documents in the "Tagging" section for general methods and systems of assigning unique or non-unique identifiers or molecular barcodes. These include:

1. US 2001/0053519 (Publication date: 2001-12-20)
2. US 2003/0152490 (Publication date: 2003-08-14)
3. US 2011/0160078 (Publication date: 2011-06-30)
4. US 6,582,908 (Issue date: 2003-06-24)

These documents, all published or issued before the priority date of US11149306, are relevant prior art.

Combination of Prior Art References and Motivation for Combination

A PHOSITA, skilled in molecular biology, genomics, and bioinformatics, would have been motivated to combine the teachings of established molecular barcoding techniques with the general knowledge of DNA structure and the known inefficiencies in next-generation sequencing (NGS) workflows to address the problem of accurately quantifying original DNA molecules.

Primary Combination: US 6,582,908 (Lander et al.) + General Knowledge of DNA Duplexes and Sequencing Inefficiencies

• US 6,582,908 (Lander et al.): This patent teaches methods for attaching unique tags (molecular barcodes) to individual nucleic acid molecules prior to amplification and sequencing. The primary purpose is to distinguish individual starting molecules, reduce errors introduced during amplification (PCR bias), and enable accurate quantification of original molecules.
• General Knowledge of DNA Duplexes: A PHOSITA would be well aware that genomic DNA exists as double-stranded molecules, each comprising two complementary strands (Watson and Crick).
• General Knowledge of Sequencing Inefficiencies: It was well-known in the art that DNA library preparation and sequencing workflows are not 100% efficient. This means that a significant portion of original DNA molecules might not be successfully ligated with adapters, amplified, or sequenced, or only one of their two complementary strands might be detected. This inherent inefficiency leads to an underestimation of the true number of original molecules in a sample.

Motivation to Combine:
A PHOSITA, seeking to improve the quantitative accuracy of DNA sequencing assays—a goal directly addressed by Lander et al. through molecular barcoding—would be acutely aware of the problem of incomplete molecular recovery and detection, especially when dealing with low-input or rare molecules (e.g., cfDNA, rare variants). Recognizing that DNA is double-stranded, and that existing barcoding methods primarily focused on tracking individual single-stranded amplicons derived from original molecules, the PHOSITA would be motivated to devise a strategy that specifically accounts for the fate of both complementary strands of an original double-stranded molecule.

If a molecular barcoding system could differentiate and track each strand of an original duplex, then observing whether both strands ("Pairs") or only one ("Singlets") were successfully processed would provide critical information about the conversion efficiency and detection probability of those original molecules. This information could then be used to infer the presence of original molecules for which neither strand was detected, thereby yielding a more accurate estimation of the total number of original double-stranded DNA fragments in the sample. This systematic approach would be a logical step for a PHOSITA aiming to maximize quantitative precision in the face of known experimental losses.

How this Combination Renders Claims Obvious:

Many claims in US11149306, particularly those related to the core "Pairs" and "Singlets" methodology, would be rendered obvious by this combination:

• Claiming "duplex tags, wherein each duplex tag differently tags the first and second complementary strands": Lander et al. teaches tagging individual molecules. Designing an adapter (e.g., Y-shaped, as mentioned in US11149306, which was a known adapter type for NGS library preparation) with molecular barcodes positioned or designed such that the tags on the resulting complementary strands could be distinguished would be a straightforward engineering choice for a PHOSITA motivated to track individual strands. This could involve, for instance, placing distinct barcodes on the two arms of a Y-adapter or using directional information inherent in the adapter design.
• Claiming "sorting sequence reads into paired reads and unpaired reads": Once complementary strands are tagged and distinguishable, classifying their sequence reads into "paired" (both detected) or "unpaired" (only one detected) categories is a routine bioinformatics task.
• Claiming "determining quantitative measures of (i) the paired reads and (ii) the unpaired reads that map to each of one or more genetic loci": Quantifying reads mapping to loci is fundamental to NGS analysis. Applying this quantification specifically to the categorized "paired" and "unpaired" reads directly follows from the motivation to track strand fate.
• Claiming "estimating... a quantitative measure of total double-stranded polynucleotide molecules... based on the quantitative measure of paired reads and unpaired reads mapping to each locus": This inferential step is the logical conclusion of the "Pairs" and "Singlets" analysis. Given the observed frequencies of detecting both strands, one strand, or inferring neither, a PHOSITA would apply standard statistical models (e.g., binomial distribution, as noted in US11149306) to estimate the overall conversion/detection efficiency and, consequently, the number of original molecules that went entirely undetected. The patent itself lists common statistical distributions (binomial, exponential, beta, or empirical distribution) as quantitative measures, indicating these are known tools for such inference.
• Claiming improved CNV or genetic variant detection: The ultimate goal of improving quantitative accuracy through the "Pairs" and "Singlets" method is to enhance downstream analyses like CNV and rare variant detection, which were known applications for molecular barcoding (as generally taught by Lander et al. and further emphasized by the need for high sensitivity in cfDNA applications, e.g., Lo et al.).

Secondary Combination: US 2011/0160078 (Lo et al.) + US 6,582,908 (Lander et al.) + General Knowledge

• US 2011/0160078 (Lo et al.): This patent focuses on highly sensitive and specific detection of rare genetic variants, particularly in cell-free nucleic acid samples. It emphasizes the need for methods that can overcome technical challenges to accurately identify low-frequency mutations.
• US 6,582,908 (Lander et al.): As described, teaches the use of molecular barcodes for accurate quantification and error reduction.
• General Knowledge: Understanding of DNA duplexes and inefficiencies in molecular biology workflows.

Motivation to Combine:
Lo et al. highlights the critical need for robust methods to detect rare genetic variants, where even small inaccuracies can lead to false positives or negatives. A PHOSITA, addressing the specific challenges of rare variant detection as emphasized by Lo et al., would recognize that existing molecular barcoding techniques (Lander et al.) could be further refined to improve quantitative accuracy. The desire for enhanced sensitivity and specificity for rare variant detection would strongly motivate the PHOSITA to account for all original molecules, including those partially or completely lost during processing. This motivation would naturally lead to considering the fate of both strands of a DNA duplex and implementing the "Pairs" and "Singlets" analysis to obtain the most accurate starting molecule count possible. The computer-implemented aspects of the claims (grouping reads, merging into consensus sequences, calling variants) are standard bioinformatics procedures for NGS data and would be readily applied in this context.

In conclusion, the inventive methods and systems disclosed in US11149306, particularly the use of duplex tags to distinguish complementary strands and the subsequent "Pairs" and "Singlets" analysis to infer unseen molecules for improved quantitative accuracy, would have been obvious to a PHOSITA by combining established molecular barcoding techniques (such as those taught by Lander et al.) with the general knowledge of DNA's double-stranded nature and the known inefficiencies of sequencing workflows, driven by the strong motivation to improve the accuracy and sensitivity of genetic variant detection (as exemplified by the objectives in Lo et al.).