Chemical Array Signal Detrending via Surface Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chemical array data analysis is hindered by systematic biases and distortions due to physical limitations and manufacturing processes, such as gradient effects, which complicate signal interpretation and data validation across arrays.
Innovation Solution
The method involves inputting signal intensity values from chemical arrays, filtering out saturated and non-uniform features, calculating log transforms, and using surface approximation to normalize and de-trend signal data, thereby correcting for trends and improving data reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gradient effects and systematic biases are present in array data, then signal intensity values can be obtained, but data accuracy and reliability deteriorate due to distortions in expression values
Solution Approach 1:
The patent extracts and removes gradient effects from the array data by fitting a surface to the signal intensity values and subtracting the fitted surface from the original data. This separates the harmful gradient component from the true biological signal, thereby improving data accuracy while eliminating systematic biases
Solution Approach 2:
The patent introduces an intermediate surface fitting step that models the gradient effects. By using this intermediate mathematical representation of the gradient, the system can systematically remove the distortion without directly manipulating the raw signal values, thus improving reliability while accounting for the harmful gradient effects
2Ease of operation
If signal intensity values are used directly for analysis, then data processing is simpler, but intra-array and inter-array comparisons become unreliable due to systematic biases
Solution Approach 1:
The patent performs detrending as a preliminary step before any comparative analysis. By removing gradient effects upfront through surface fitting and subtraction, the data is pre-conditioned for reliable comparisons, maintaining ease of operation while ensuring that subsequent intra-array and inter-array comparisons are trustworthy
3Reliability
If log transform and surface approximation are applied to remove trends, then data reliability improves, but processing complexity increases
Solution Approach 1:
The patent transforms the data from linear signal intensity space to log-transformed space before applying surface approximation. This parameter change linearizes the relationship between gradient effects and signal values, making the surface fitting more effective at removing trends while maintaining computational feasibility, thus improving reliability without excessive complexity
Data Source
AI summary
Methods, systems and computer readable media for removing trends in signal intensity values from features on a chemical array. Inputted signal values from features on the array are surface fitted to calculate a surface approximation. The surface approximation is normalized and used to de-trend the signal intensity values from the features.