Searching for Drug Synergy in Complex Dose-Response Landscapes Using an Interaction Potency Model.
Yadav B, Wennerberg K, Aittokallio T, Tang J. · 2015
Read Searching for Drug Synergy in Complex Dose-Response Landscapes Using an Interaction Potency Model. by Yadav B, Wennerberg K, Aittokallio T, Tang J. (2015) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4759128.txt.
Introduction
Bhagwan Yadav, Krister Wennerberg, Tero Aittokallio, and Jing Tang propose a new scoring model in Computational and Structural Biotechnology Journal in 2015. The model judges whether two drugs help or hinder each other when given together. Such pairs matter for stubborn diseases, but older scoring rules break down on the rich dose grids that modern screens produce.
Their model, called zero interaction potency, compares how a drug's strength curve shifts when its partner joins. A single score marks the gap from the no-interaction case across the dose grid. On a large cancer drug screen, the score caught known strong pairs while keeping false alarms low. The team also maps the full grid, so users can see which dose zones help and which hurt.
Catalog record
| CREATOR | Yadav B, Wennerberg K, Aittokallio T, Tang J. |
|---|---|
| YEAR | 2015 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Searching for Drug Synergy in Complex Dose–Response Landscapes Using an Interaction Potency Model By Bhagwan Yadav, Krister Wennerberg, Tero Aittokallio, Jing Tang Published in: Computational and Structural Biotechnology Journal DOI: 10.1016/j.csbj.2015.09.001 ABSTRACT Rational design of multi-targeted drug combinations is a promising strategy to tackle the drug resistance problem for many complex disorders. A drug combination is usually classified as synergistic or antagonistic, depending on the deviation of the observed combination response from the expected effect calculated based on a reference model of non-interaction. The existing reference models were proposed originally for low-throughput drug combination experiments, which make the model assumptions often incompatible with the complex drug interaction patterns across various dose pairs that are typically observed in large-scale dose–response matrix experiments. To address these limitations, we proposed a novel reference model, named zero interaction potency (ZIP), which captures the drug interaction relationships by comparing the change in the potency of the dose–response curves between individual drugs and their combinations. We utilized a delta score to quantify the deviation from the expectation of zero interaction, and proved that a delta score value of zero implies both probabilistic independence and dose additivity. Using data from a large-scale anticancer drug combination experiment, we demonstrated empirically how the ZIP scoring approach captures the experimentally confirmed drug synergy while keeping the false positive rate at a low level. Further, rather than relying on a single parameter to assess drug interaction, we proposed the use of an interaction landscape over the full dose–response matrix to identify and quantify synergistic and antagonistic dose regions. The interaction landscape offers an increased power to differentiate between various classes of drug combinations, and may therefore provide an improved means for understanding their mechanisms of action toward clinical translation. INTRODUCTION Drug combinations that consist of multiple chemical agents have shown great promises to improve efficacy and overcome resistance for treating complex and refractory diseases. In cancer therapy, for example, an effective drug combination may target multiple proteins or pathways that are aberrantly activated in disease, but not in normal cells, and thus reduces the chances that a cancer will develop drug resistance by activating bypassing pathways, when compared to standard monotherapies [1], [2], [3]. To accelerate the discovery of novel drug combinations using an empirical approach, preclinical drug screening platforms have been developed that allow for assaying the phenotypic endpoint responses in vitro or ex vivo. With high-throughput techniques, it has become possible to systematically evaluate the pairwise combinations from a large collections of both approved and investigational chemical compounds [4]. Such functional screening approaches have increasingly led to advances in the characterization of drug–drug interactions for multiple diseases, and prioritization of effective and safe drug combinations for further clinical examination. To quantify the interaction between drugs, the observed combination response is often compared to the expected effect under the assumption of non-interaction predicted by a reference model. When the combination response is greater than what is expected, then the combinations is classified as synergistic, while antagonism is concluded when the combination produces less than the expected effect. Currently, there are three popular classes of reference models: Highest single agent (HSA) model [5], Loewe additivity model [6] and Bliss independence model [7]. These reference models, together with many of their subsequent variants and extensions, have been developed based on different assumptions about the expected effect of non-interaction. The HSA model, or Gaddum's non-interaction model, assumes that the expected combination effect equals to the higher individual drug effect at the dose in the combination, representing the idea that a synergistic drug combination should produce additional benefits on top of what its components can achieve alone. In many preclinical drug combination studies, however, even a drug combined with itself can easily produce an excess over HSA. For more stringent synergy classification, the Loewe additivity and Bliss independence models are being widely used. The Loewe additivity model defines the expected effect as if a drug was combined with itself, while the Bliss independence model utilizes probabilistic theory to model the effects of individual drugs in a combination as independent yet competing events. Due to the inherent differences in the model assumptions, there is a lack of consensus on which references model one should use in an unbiased and statistically robust manner. As pointed out by many others [8], [9], [10], there is still no standardized guideline on how to choose the optimal reference model. Apart from the disagreement among the theoretical aspects, there are considerable pragmatic challenges on how to apply the Loewe and Bliss models for analyzing drug combination experiments. First, the currently available scoring tools are rather limited and suboptimal for a high-throughput setting. For instance, CompuSyn is one of the most-cited standalone software packages used to calculate the interaction score based on the Loewe model [11]. However, it only allows for a manual input of dose–response data for one drug combination at a time, and thus becomes quickly unfeasible for analyzing large-scale drug screening data which typically consists of hundreds of combinations. For implementations of the Loewe model, there are a few packages in R, such as drc [12] and SYNERGY [13], but these packages are not flexible enough to handle the noise in a typical high-throughput drug screen, which tend to invalidate many of the widely used curve fitting functions. As a result, drugs that have negative or irregular responses may easily lead to a fitting error that breaks down the program. Secondly, many existing methods utilize a summary interaction score to characterize the overall drug interaction effect, which may be sufficient for the initial filtering of potential synergistic drug combinations. However, a follow-up confirmatory screen often utilizes a dose–response matrix for which a single interaction parameter cannot effectively capture the synergies and antagonisms that may occur within specific dose regions only. With an increasing number of drug combinations that have been tested using the dose–response matrix design, there is a critical need to develop efficient and robust computational tools to systematically evaluate and visualize the drug interaction at the level of individual dose combinations. To address these challenges, we introduced a novel reference model, named zero interaction potency (ZIP), which overcomes many of the limitations of the existing models. By combining the advantages of both the Loewe and Bliss models, the ZIP model assumes that two non-interacting drugs are expected to incur minimal changes in their dose–response curves. We calculated a delta score to quantify the deviation from the expectation of ZIP for a given dose pair and utilized the average delta over a dose–response matrix as a summary interaction score for a drug combination. To test this model, we analyzed the data from a recent high-throughput drug combination study in cancer [14]. We showed that the ZIP-based delta scoring approach tolerates well the experimental noise in such combination screens and provides an improved solution for identifying true synergistic interactions, while keeping the false positive rate relatively low. Further, we demonstrated how an interaction landscape visualization of delta scores can facilitate the identification of clinically relevant drug synergy within specific dose regions. We believe our drug interaction scoring method has the potential to allow for a systematic evaluation of drug combinations in a high-throughput setting, and may therefore become