Related Experiment Videos
Predicting behavior of an enzyme-linked immunoassay model by using commercially available neural network software
1West Penn Center for Neuro-oncology, Western Pennsylvania Hospital, Pittsburgh 15224-1722.
Clinical Chemistry
|December 1, 1993
Summary
Neural networks offer a cost-effective solution for optimizing immunoassays like ELISA. This study demonstrates how neural networks can predict assay performance, saving time and reagents in laboratory settings.
Area of Science:
- Biotechnology
- Computational Biology
- Immunology
Background:
- Immunoassay development, particularly ELISA, is often time-consuming and resource-intensive.
- Neural networks represent a novel multivariate analysis approach with potential for optimizing laboratory workflows.
Purpose of the Study:
- To evaluate the application of neural networks for optimizing enzyme-linked immunosorbent assays (ELISA).
- To assess the potential for cost and time savings in immunoassay development using computational methods.
Main Methods:
- A feed-forward neural network was trained using the Brainmaker software package.
- Four variable ELISA conditions (antigen concentration, antibody titers, and development time) were used as inputs.
- The network predicted measured absorbances in a model human serum albumin ELISA.
Main Results:
- The neural network successfully predicted the impact of input variables on absorbance.
- The model demonstrated adequate prediction accuracy for ELISA performance.
Conclusions:
- Neural networks can effectively optimize ELISA conditions computationally.
- This approach offers significant savings in reagents and technician time compared to traditional laboratory optimization.