Hi everyone!
Recently I was doing a lot screening of drug pairs in 5 osteosarcoma cell lines using this amazing website https://synergyfinder.fimm.fi. I am puzzled why a lot of combinations have positive HSA and Bliss scores, but negative Loewe. Only few combinations have all three scores positive. How do I interpret the results? Am allowed to consider a drug combination synergistic when HSA and Bliss scores are positive, but Loewe score is negative?
Looking forward to your comments
Hi Konstantin,
Sorry for late reply. "Which references model should we use when assessing interaction between two drugs" is a very difficult question on which many efforts have been done Article What is synergy? The Saariselkä agreement revisited
. However, due to the inherent differences in the model assumptions, there is still a lack of consensus on reference model selection.Understanding the underlying hypothesis might be helpful to chose the optimal model for individual project. Details about four reference models (HSA, Loewe, Bliss and ZIP) used by SynergyFinder could be found on the “Reference models” section of our user guide page or from this article: Article Searching for Drug Synergy in Complex Dose-Response Landscap...
.For interpreting the value of synergy scores, as SynergyFinder has normalized input data as percentage inhibition, they can be directly interpreted as the proportion of cellular responses that can be attributed to the drug interactions. (e.g., synergy score 20 corresponds to 20% of response beyond expectation). According to our experiences, the synergy scores near 0 gives limited confidence on synergy or antagonism. So when synergy score:
· Less than -10: we consider the interaction between two drugs is likely to be antagonism;
· From -10 to 10: we consider the interaction between two drugs is likely to be additivity;
· Larger than 10: we consider the interaction between two drugs is likely to be synergy.
Hope these could answer your questions.
Finally, thanks for using SynergyFinder. We have developed a new data portal for drug combination called “DrugComb” (https://drugcomb.fimm.fi/) which is also led by professor Jing Tang. It provides synergy score and sensitivity scores calculation for throughput drug combination data as well as other features for facilitating drug combination exploration. Welcome to have a try and give us some comments.
Hi Konstantin,
Sorry for late reply. "Which references model should we use when assessing interaction between two drugs" is a very difficult question on which many efforts have been done Article What is synergy? The Saariselkä agreement revisited
. However, due to the inherent differences in the model assumptions, there is still a lack of consensus on reference model selection.Understanding the underlying hypothesis might be helpful to chose the optimal model for individual project. Details about four reference models (HSA, Loewe, Bliss and ZIP) used by SynergyFinder could be found on the “Reference models” section of our user guide page or from this article: Article Searching for Drug Synergy in Complex Dose-Response Landscap...
.For interpreting the value of synergy scores, as SynergyFinder has normalized input data as percentage inhibition, they can be directly interpreted as the proportion of cellular responses that can be attributed to the drug interactions. (e.g., synergy score 20 corresponds to 20% of response beyond expectation). According to our experiences, the synergy scores near 0 gives limited confidence on synergy or antagonism. So when synergy score:
· Less than -10: we consider the interaction between two drugs is likely to be antagonism;
· From -10 to 10: we consider the interaction between two drugs is likely to be additivity;
· Larger than 10: we consider the interaction between two drugs is likely to be synergy.
Hope these could answer your questions.
Finally, thanks for using SynergyFinder. We have developed a new data portal for drug combination called “DrugComb” (https://drugcomb.fimm.fi/) which is also led by professor Jing Tang. It provides synergy score and sensitivity scores calculation for throughput drug combination data as well as other features for facilitating drug combination exploration. Welcome to have a try and give us some comments.
Hi! Thank you for this terrific answer. It is very helpful. Your website drugcomb.fimm.fi is a treasury. So many hypotheses can be tested for their viability at once.
I have moved on from osteosarcoma to myeloma. Therefore, I work a lot with different myeloma cell lines, including RPMI-8226, which has a lot of entries on your drugcomb website. Sorting by descending HSA or Bliss score for RPMI-8226 one immediately comes to highly synergistic combos including different TKIs. And it has a plausable explanation, since RPMI-8226 is positive for a multidrug efflux pump ABCG2, which is likely to be inhibited by TKIs like lapatinib, nilotinib and others.
As for drugcomb website, I wish you could sort the entities in more details. Not just haematopoietic and lymphoid, but myeloma, AML, lymphoma, ALL, CML, etc. Also, the Advanced search option might require a different design. I have clicked on "Remove" instead of "Search" too many times.
In general, the drugcomb website is very helpful. Keep doing the great work)))
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line