AI Base Calling for Overlapping Sequencing Clusters

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing nucleic acid sequencing technologies face limitations in resolving data from closely proximate or spatially overlapping clusters, leading to reduced throughput and increased computational resource requirements for base calling.

Innovation Solution

An artificial intelligence-based method using deep neural networks, specifically convolutional and recurrent neural networks, processes sequencing images to segregate and base call nucleotides by employing spatial and temporal convolution layers to handle misalignment and sequence-specific context, enabling many-to-many base calling for improved resolution and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional base calling methods are used to process sequencing images, then the computational resource requirements increase and processing time increases, but the resolution and quality of sequencing data decrease

Engineering Contradiction:
Improvesequencing data qualityVSAvoidcomputational processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent transforms the base calling problem by changing the parameter representation from traditional one-to-one mapping to many-to-many relationships. The neural network learns to associate multiple image patches with multiple base calls, enabling parallel processing that reduces computational time while maintaining or improving sequencing data quality through pattern recognition across multiple sequences simultaneously

Inventive Principle:
Principle #35Parameter changes

2Productivity

If traditional base calling methods are used to process sequencing images, then the computational resource requirements increase, but the throughput decreases

Engineering Contradiction:
Improvesequencing throughputVSAvoidcomputational resource requirements
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent merges the processing of multiple sequencing images and multiple base calling tasks into a single unified neural network operation. By combining multiple image patches and their corresponding base calls into one training and inference operation, the system achieves higher throughput while managing computational resources more efficiently through shared feature extraction and parameter learning

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If deep neural networks with many-to-many base calling are used, then the throughput increases and computational efficiency improves, but the algorithm complexity increases

Engineering Contradiction:
Improvesequencing throughputVSAvoidalgorithm complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the complex many-to-many base calling problem into manageable components: image patch extraction, feature encoding through convolutional layers, sequence generation through recurrent layers, and base calling prediction. This segmentation allows the complex algorithm to be implemented as a modular neural network architecture that can be trained end-to-end while maintaining computational efficiency and enabling higher throughput

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4107735B1Artificial intelligence-based many-to-many base calling
Publication Date: 2025.11.26 ILLUMINA INC
  • EP4107735B1 patent drawingFigure 1A~1B
  • EP4107735B1 patent drawingFigure 1C
  • EP4107735B1 patent drawingFigure 1D

AI summary

The technology disclosed relates to artificial intelligence-based base calling. The technology disclosed relates to accessing a progression of per-cycle analyte channel sets generated for sequencing cycles of a sequencing run, processing, through a neural network-based base caller (NNBC), windows of per-cycle analyte channel sets in the progression for the windows of sequencing cycles of the sequencing run such that the NNBC processes a subject window of per-cycle analyte channel sets in the progression for the subject window of sequencing cycles of the sequencing run and generates provisional base call predictions for three or more sequencing cycles in the subject window of sequencing cycles, from multiple windows in which a particular sequencing cycle appeared at different positions, using the NNBC to generate provisional base call predictions for the particular sequencing cycle, and determining a base call for the particular sequencing cycle based on the plurality of base call predictions.