Over the past month I have been exploring DistilBio to look at specific use cases that would be useful to someone working in the area of drug discovery. I came up with several interesting and encouraging results. In this post, the first in this series, I will show you how DistilBio can be used to search for the target of a drug.
The drug I have chosen is Sitagliptin (Trade Name: Januvia), an anti-diabetic drug developed and marketed by Merck & Co. Lets see what DistilBio tells me about this drug.
First, lets look at what is Sitagliptin and the associated disease indication.
Query: Sitagliptin > disease indication
Below, we can see the result..
We can see that Sitagliptin is a hypoglycemic/anti-diabetic drug. The next question that comes to mind is what are the protein targets of Sitagliptin? Lets also add a node to see the function of the protein targets.
Query: Sitagliptin > protein > function
and the result is
We can see 2 protein targets DPP4 and SO4C1, both in humans. (Uniprot Protein Identifiers are displayed in DistilBio). Highlighting the column for protein targets allows us to view information about the proteins.
The functions show us that DPP4 is a dipeptidyl peptidase enzyme that cleaves peptides to regulate their levels and also plays a role in T-cell activation.
SO4C1 is a transporter of pharmacological substances. So I am assuming that DPP4 is the primary protein target of Sitagliptin.
Now this is puzzling. What role does DPP4 play in diabetes or regulating blood glucose? Will proteins interacting with DPP4 give us this information? Lets also focus on protein interactions in humans.
Query: Sitagliptin > protein > protein > human
We get the following results..
We can see a list of proteins that interact with DPP4. But which of these are involved in blood glucose regulation?
The next step is to filter these to find the proteins that interact with DPP4 and also play a role in glucose level regulation. To do this, select the column with the interacting proteins and search for “glucose” in the field named ‘filter’.
That’s interesting. We get 2 proteins that interact with DPP4 – GIP and GLUC. Further exploration of the data for GIP and GLP-1 (a polypeptide product of Glucagon) indicate that these proteins play an important role in insulin secretion.
Bingo!! … Using DistilBio we were able to figure out that Sitagliptin, an anti-diabetic drug targets/inhibits the protein DPP4. DPP4 interacts with Gastric Inhibitory Protein (GIP) and Glucagon like peptide -1 (GLP-1). GIP and GLP-1 are potent stimulators of insulin and also suppress glucagon. Currently, DistilBio does not give us the explicit type of interaction (activation or inhibition) between DPP4 and GIP/GLP-1, but I am assuming that DPP4 inhibits GIP and GLP-1, therefore inhibiting DPP4 activates these 2 proteins. Types of interactions between proteins, genes, drugs etc., is a feature that will be available in DistilBio soon.
Here, I have shown the use of DistilBio to find the targets of a drug and its role in a disease indication. Similarly, you could use it to find common targets of drugs, drug interactions, protein interactions etc., which would be a useful tool for drug discovery.
Over the next several weeks, I will be exploring some more practical examples of how DistilBio could be used. Stay tuned for more use cases.
About the Author: Preethi is currently with Metaome as a Consultant Biologist. A post-graduate in Biochemistry and Bioinformatics, she has worked in Systems Biology and computational modeling of biological processes.
Very succinct! I guess you could identify possible side effects as well by looking at the list of proteins that interact with DPP4.
Very interesting example. Nicely illustrated. I can tell that DisTil is going to be a powerful tool in collating biological information for bench biologists like myself. I will be looking forward to seeing more examples. Can you give us any examples on protein-protein interactions and what kind of information I can obtain for a given protein? Great work!
Disclaimer: This comment is written in a positively critical point of view, and therefore it would be best to avoid a negative spin on it.
Quite interesting! But, I would not be surprised to find that DistilBio’s USP is not as unique as it seems. Actually, by your own admission, you have inadvertently pointed towards what the USP should be – rendering information about the the explicit type of protein-protein interaction.
Today, in drug discovery, particularly metabolic diseases such as diabetes, several metabolic enzymes are the “in thing” as drug targets. Now, this is a completely different domain from simple protein-protein interactions. Enzyme-enzyme interactions tend to have a much broader scope and often times knowledge of these interactions holds the key to a successful drug design in the metabolic realm. So, if I were to be in your shoes, I’d be thinking very hard about the possibilities of defining this particular problem and finding an effective solution that would guarantee launching DistilBio to the top of the charts, in any biochemists’ and drug designer’s eyes.
I wish you good luck with this interesting piece of work and hope to see DistilBio do well in near future!
Amar, thank you for your feedback. The initial focus of focus of DistilBio was breadth rather than depth. However your point is well taken. We are prioritizing areas to drill down into and being able to point to explicit protein protein interactions is definitely an important one.
Please do keep looking at DistilBio as we continue to add data as well as features. We would love to hear from you.
Pingback: Metaome Helps Bench Biologists Get More Value From Linked Data - semanticweb.com
If this turns out to be useful we would need to have similar services implemented by the NCBI or EBI or other completely entities that make all of their code and underlying data publicly available. I am not saying that business development in this area is not useful or desirable but the only truly persistent and useful services are completely open.