Help page

What is Target Fisher?

Target Fisher is a reverse docking tool. It searches a pool of targets trying to find one where a given molecule could bind.

How does Target Fisher work?

Target Fisher first does a regular docking with the selected targets and the input molecule. Then, using machine learning methods, evaluates the binding poses provided by the docking to predict a possible activity.

Exactly what do you mean by machine learning?

Machine learning is a broad field with a grand collection of methods. Particularly, we use support vector machines (SVM) to create a binding predictor. This predictor takes a target/ligand complex and produces a numerical predictor of activity. We then compare this predictor with the ones obtained for the training dataset to classify the molecule as either "inactive/inconclusive/potentially active". The classification is based on wether the predictor falls in the quartiles Q₁/Q₂-Q₃/Q₄, respectively.

Are there any restrictions to the use webserver use?

Yes, you can only have one submission queued with up to 5 molecules at any time. Our server is not the best, and its also shared with other services. If you found it useful, please let us know so we can ask for cold hard cash in the form of a grant.

Is this a web-only app?

No, you can get Target Fisher as a desktop application (for linux). Go to https://gitlab.com/gqc/target-fisher, install it and you can use it freely. Its the same as the one offered in this page.

Can I take a look at the code?

Yes, check the gitlab link above and take a look. If you are into software development and would like to contribute, any help is welcomed. We are chemists, so we know our code is not the best in the whole wide world.

Do you take responsibility if compound data is somehow leaked?

No.