In taylo5jm/interactions: Classification of gene expression profile sets generated from combination stimuli

We discuss interactions, a R package with functions to classify an interaction profile (set of gene expression outputs) into an interaction class and interaction mode, a well-defined behavior. This theoretical framework is discussed in Combinatorial code governing cellular responses to complex stimuli (Cappuccio et. al).

To summarize the aforementioned paper, consider the a set of 4 gene expression outputs. Signals: $\theta{}$, $X$, $Y$, $X + Y$. We are interested in predicting the effects of $X + Y$ that cannot be predicted from signal $X$ or signal $Y$ alone. The set of 4 gene expression outputs can result in $82$ interaction profiles, or statistically different interaction profiles.

These interaction profiles can be further classified into interaction classes. An interaction class is defined by: $$\Delta{X} + \Delta{Y} > \Delta{X + Y}$$ Positive $$\Delta{X} + \Delta{Y} < \Delta{X + Y}$$ Negative

where the fractional effects of any signal can be calculated by subtracting the gene expression output of \theta{} from the gene expression output of the signal. Interaction profiles that do not satisfy one of these conditions are labeled as additive or null.

Additive interaction profiles are defined by: $$\Delta{X} +\Delta{Y} = \Delta{X + Y}$$

Null interaction profiles are defined as: $$\Delta{\theta{}} = \Delta{X} = \Delta{Y} = \Delta{X + Y}$$

Interactions classified into a class can then be classified by an interaction mode with a well-defined mathematical behavior. The positive and negative interaction classes are comprised of 5 modes each. An example of a mode in the positive class would be emergent positive synergy. Each mode in a class has a conjugate mode in the opposite class (e.g. emergent negative synergy).

taylo5jm/interactions documentation built on May 31, 2019, 3:57 a.m.