Target-Binding Oligonucleotides With Intentional Mismatches for Polymorphisms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems are inefficient in generating target-binding oligonucleotides that effectively target nucleic acid sequences with polymorphisms due to the inability to introduce mismatches and asymmetry, relying on human assessments that fail to account for subtle variations, and lack the capability to generate multi-target-binding oligonucleotides.
Innovation Solution
Employing machine learning systems, such as neural networks and evolutionary algorithms, to process target nucleic acid sequences and generate engineered target-binding oligonucleotides with intentional mismatches and asymmetry, enhancing their ability to target sequences with polymorphisms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional systems generate target-binding oligonucleotides that fully match a particular target nucleic acid sequence, then the oligonucleotide can be designed to target a specific sequence, but it cannot effectively target nucleic acid sequences with polymorphisms
Solution Approach 1:
The patent introduces asymmetry by deliberately designing oligonucleotides with mismatches at specific positions rather than achieving perfect matches. This asymmetric design allows the oligonucleotide to bind to multiple target sequences with different polymorphisms while maintaining sufficient binding affinity through the mismatch penalty mechanism.
Solution Approach 2:
The system changes the matching parameter from perfect match (identity = 100%) to intentional mismatch (identity < 100%). By controlling the mismatch parameter at specific positions, the system enables the oligonucleotide to target sequences with polymorphisms while maintaining reliable binding through the designed mismatch penalty.
2Ease of manufacture
If human assessments are used to configure target-binding oligonucleotides, then the process can be performed with current technology, but the system cannot introduce mismatches to effectively target nucleic acid sequences with polymorphisms
Solution Approach 1:
The patent replaces human assessment mechanisms with machine learning systems that can automatically analyze sequence data and identify subtle variations. The ML system processes large datasets of target sequences and polymorphisms to determine optimal mismatch positions, eliminating the limitations of human assessment in detecting subtle variations.
Solution Approach 2:
The system creates computational models that replicate and extend human expertise by learning from extensive training data. The ML algorithms copy the pattern recognition capabilities of human experts while adding the ability to process vast amounts of data and identify patterns that would be imperceptible to humans.
3Device complexity
If conventional systems are used to generate target-binding oligonucleotides, then the system structure remains simple, but it cannot generate multi-target-binding oligonucleotides by introducing mismatches
Solution Approach 1:
The patent implements a universal ML-based oligonucleotide design system that can handle multiple target sequences and polymorphism patterns with a single integrated framework. The system takes as input any set of target sequences and automatically generates appropriate oligonucleotides with mismatches, eliminating the need for separate design approaches for different target types.
Solution Approach 2:
The system segments the oligonucleotide design process into distinct functional components: sequence input processing, polymorphism identification, mismatch position determination, and oligonucleotide generation. This segmentation allows each component to be optimized independently while working together to achieve multi-target binding capability.
Data Source
AI summary
Current machine learning methods for target binding oligonucleotides design are limited to considering natural sequences in the targets. Here, Applicants generated novel target binding oligonucleotides—with multiple mismatches to any natural sequence—that are optimized for desired properties. These novel target binding oligonucleotides offer more sensitive and specific detection of, for example, pathogen genome variation than baseline design methods, and they illuminate a new, interpretable design rule that broadens nucleic acid sequence targeting.