Modeling Cellular Responses in RAW 264.7 Cells
Author Information
Author(s): Wu Yang, Johnson Gary L, Gomez Shawn M
Primary Institution: University of North Carolina School of Medicine
Hypothesis
Can data-driven modeling improve our understanding of cellular signaling and output responses?
Conclusion
Multivariate modeling approaches can effectively predict cytokine responses in RAW 264.7 cells, although the utility of time-dependent metrics is limited.
Supporting Evidence
- Model R2 values ranged from 0.48 to 0.93 for predicting cytokine responses.
- PLS models required fewer variables than PCR models to achieve similar prediction accuracy.
- Time-dependent metrics were found to be of mixed value in improving predictive performance.
Takeaway
Scientists used computer models to understand how cells respond to signals, but they found that having more time points in their experiments could help them make better predictions.
Methodology
The study used partial least squares (PLS) and principal components regression (PCR) to analyze signaling data from RAW 264.7 cells.
Potential Biases
Potential bias due to unmatched stimulant concentrations across experiments.
Limitations
The study was limited by the low number of matched input conditions and the sparsity of time points in the signaling data.
Participant Demographics
RAW 264.7 macrophage-like cells were used in the experiments.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website