Table of content:


What is the DASPfind tool?

DASPfind is a tool that aims to predict new Drug-Target interactions from a network that encodes information about the known Drug-Target interactions, similarities between the drugs and similarities between targets. Our results are presented in the following paper:

DASPfind: New Efficient Method to Predict Drug-Target Interactions
Wail Ba-alawi, Othman Soufan, Magbubah Essack, Panos Kalnis, and Vladimir B. Bajic
Cheminformatics, DOI: 10.1186/s13321-016-0128-4

How does DASPfind work?

Our algorithm for predicting new drug-target interactions is based on all simple paths of particular lengths on such network. The main idea in our method is to utilize the similarity information within the sub networks and combine it with information from the topology of the heterogeneous network to predict and rank new interaction edges.

How to use this web tool?

Upload Drug-Target Interactions Dataset: An NxM matrix tab delimited file where rows are drugs and columns are targets. The values of the matrix elements are in the range [0,1], where 0 means no known interaction and 1 is a known interaction.

Upload Drug Similarity Dataset: An NxN matrix tab delimited file where rows and columns are drugs and the values represent the similarities between the drugs. The values of the matrix elements are in the range [0,1], where 0 means no similarity between the two drugs at all and the larger the value the more similar are the drugs.

Upload Target Similarity Dataset: An MxM matrix tab delimited file where rows and columns are targets and the values represent the similarities between the targets. The values of the matrix elements are in the range [0,1], where 0 means no similarity between the two targets at all and the larger the value the more similar are the targets.

Drug Similarity Threshold: Threshold to constraint the drugs similarities to be above it. The values range for this threshold is between (0,1).

Target Similarity Threshold: Threshold to constraint the targets similarities to be above it. The values range for this threshold is between (0,1).

Email address: The user's email address for sending back the results.

How does the results look like?

After submitting the job, the user will receive an email confirming the job submission with an ID for reference purpose. Once the job is completed, the user will receive an email with an attached results file. This file contains an NxM matrix, where rows are drugs and columns are targets. This matrix is of the same size and order as the original drug-target interactions file submitted by the user. However, now the values of the elements in the matrix represent the raw scores of our method DASPfind. For each drug (row), if the user sort the columns by value, he will get a ranked list of possible targets where the higher the value, the more confident and reliable is the prediction. The results in our paper shows that DASPfind reports very reliable results for each drug when considering the top 1,2 and 5 ranked predictions.

How long it will take to run DASPfind?

The execution time depends on the size of the datasets and how busy is our dedicated server. For example, Nuclear Receptor Dataset in our paper takes few seconds to finish predicting targets for all the drugs in that dataset. However, if the dataset is larger then it can take much longer (up to several hours) to finish predicting targets for all the drugs in that dataset. Note that our method try to find all the simple paths connecting a drug and a target under consideration. We utilize parallelization in order to speed up this process. If it takes longer than you expect to receive your results, please contact us with the job ID at: daspfind.kaust@gmail.com or wail.baalawi@kaust.edu.sa