Gerotherapy combinations: how to distinguish synergy from alchemy
Lecturers Maxim Likhter and Liza Ignatova explained how to test combinations of drugs and other interventions against aging
On 14 August, Maxim Likhter and Liza Ignatova gave a lecture on how to test combinations of interventions against aging. They recommend defining a measurable outcome in advance, along with the rule used to compare the effect of the combination.
A single drug rarely affects every process associated with age-related changes. Each pair of interventions requires decisions about dose and order of administration, while a lifespan experiment may take years. Before computational screening begins, researchers must therefore define what outcome will count as a benefit of the combination.
The first step is to choose the experimental outcome, such as the proportion of dead cells, lifespan, probability of survival, or another measure. A cell proportion has an upper limit of 100%, whereas lifespan does not. The same arithmetic therefore produces different expectations for different outcomes.
The next step is to define the expected result for a pair whose components act independently. Some comparison rules assume that the interventions act independently. Others compare the pair with the better individual intervention or account for the dose ratio. The same observation can receive different assessments under different rules. The term “synergy” therefore refers to the chosen method of comparison.
The lecture presents the experiment in the following order: a control group, each intervention administered separately, their combination, and an assessment of toxicity, meaning harm caused by the combination. The speakers recommend comparing a full dose of one intervention with two halves of that dose administered together. If this division alone produces “synergy,” the chosen comparison method is generating the result.
A statistical interaction measures deviation from the selected model. Pharmacological synergy compares the observed effect with the effect expected for the pair. A biological explanation addresses a different question: how one intervention changes the response to the other. The outcome, comparison rule, and controls must be defined first. Computational methods can then select pairs for experimental testing.