Bioinformatics assistant scientist Tongjun Gu Ph.D. has developed a two-step logistic regression model for association studies where known factors can serve as controls.
The model is helpful for studies seeking to identify novel regulators in the RNA editing process where the dominant enzyme was found in the first step and served as a control in the second step.
Advantages of the model:
- It considered the sequencing depth in the analysis.
- It can be validated in the first step by the association of a biology phenomenon with the control gene, which is normally easy to obtain. Thus it is very useful, especially when the validation was difficult to be done in the second step.
- It compares with the control gene, more confident and more functional novel regulators can be found.
The study was published in the Computers in Biology and Medicine journal in an article titled “Systematic identification of A-to-I editing associated regulators from multiple human cancers”.