Sequencing Base Calling With Deep Learning for Raw Luminescence Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine learning methods for nucleotide identification in nucleic acid sequencing require manual feature engineering, which is inefficient and often leads to suboptimal performance due to the difficulty in determining the best features for reliable nucleotide identification.
Innovation Solution
Utilizing deep learning models, specifically convolutional neural networks and connectionist temporal classification-fitted neural networks, to process sensor data directly from luminescent labels during nucleotide incorporation events, eliminating the need for manual feature selection and improving nucleotide identification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual feature engineering is used in machine learning methods for nucleotide identification, then the system is easier to implement and understand, but the performance is suboptimal and efficiency is reduced due to difficulty in determining the best features
Solution Approach 1:
The deep learning model automatically performs feature extraction and selection from raw sensor data without requiring manual intervention. The model learns informative features directly from the data during training, eliminating the need for domain experts to manually engineer features and enabling the system to adapt to different sequencing conditions autonomously
Solution Approach 2:
The patent replaces the manual mechanical process of feature engineering with an automated deep learning system. Instead of manually selecting and extracting features through computational algorithms, the deep learning model automatically learns hierarchical features from raw data, substituting human effort and manual optimization with automated neural network processing
2Productivity
If manual feature engineering is performed to identify informative features, then the process is more controllable and interpretable, but it is inefficient and leads to suboptimal performance
Solution Approach 1:
The deep learning model autonomously identifies and learns the most informative features from raw sensor data during training, without requiring manual feature selection. This self-service approach enables the system to automatically adapt to different sequencing conditions and optimize feature extraction for maximum identification accuracy
Solution Approach 2:
The patent transforms the feature representation by learning hierarchical features through multiple layers of the deep learning model. The model changes the parameters of feature extraction dynamically during training, adapting to the specific characteristics of the sequencing data and optimizing the transformation from raw sensor signals to nucleotide identification features
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The deep learning approach enhances the reliability and accuracy of nucleotide identification in nucleic acid sequencing by automatically learning informative features from detected light emissions, leading to improved sequencing performance.
Implementation Method 1
accessing data obtained from detected light emissions by luminescent labels associated with nucleotides during nucleotide incorporation events
Data Source
AI summary
A method includes obtaining, from one or more sequencing devices, raw data detected from luminescent labels associated with nucleotides during nucleotide incorporation events; and processing the raw data to perform a comparison of base calls produced by a learning enabled, automatic base calling module of the one or more sequencing devices with actual values associated with the raw data, wherein the base calls identify one or more individual nucleotides from the raw data. Based on the comparison, an update to the learning enabled, automatic base calling module is created using at least some of the obtained raw data, and the update is made available to the one or more sequencing devices.