A friendly reminder: click the title to show help.

Descriptors Generation

①Input a string with letters, numbers or underline as your unique jobname.
②Upload a ZIP file for the calculation of descriptors:
If you want to generate Target-specific (one protein and multiple ligands) descriptors, your ZIP file should contain a protein file (.pdb) and a ligand file (.sdf/.mol2).

If you want to calculate Generic (several protein-ligand complexes) descriptors, your ZIP file containing the same number of folders as the target number, and each folder needs to contain a protein file and a ligand file (one molecule).
③④Select descriptors: (default:none;requirement:select at least one tool)
There are two ways to choose descriptor generation tools:
(i)choose directly from the table on the right of the web page.
(ii)quickly select a series of descriptor generation tools using the buttons provided on ASFP.
  • All: Select all the available descriptor generation tools for calculation.
  • Cancel: Disselect all the selected checkbox.
  • Reverse: Cancel selected tools, select previously unselected tools.
  • Pocket: Choose the Dpocket for protein pocket descriptor generation
  • Energy terms: Select 6 tools to compute energy terms from existing scoring functions.
    (AffiScore, AutoDock, DSX, GalaxyDockBP2, NNScore, and SMoG2016)
  • Fingerprint: Select fingerprint generation tools including molecule fingerprint.
    ( PubchemFP and ECFP )
⑤Make sure you have followed the guide and submit the job.

AI-based scoring functions

①Input a string with letters, numbers or underline as your unique jobname.
②Upload a protein file: In PDB format. The file size should larger than 20kb.
③Upload an actives file: In MOL2 or SDF format. The molecules in this should be binders to the protein.
④Upload a decoys file: In MOL2 or SDF format. The molecules in this file should be nonbinders to the protein.
⑤Upload a test file: In MOL2 or SDF format. Upload ligands of interest to predict whether they are binders or not.
⑥Choose a machine learning algorithm: ( Default: Support Vector Machine)
Three ML algorithms are available.
[Support Vector Machine (SVM), eXtreme Gradient Boosting (XGBoost), Random Forest (RF)]
⑦Hyperparameter (optional)
Decide which hyperparameters to be optimized and the range of these hyperparameters.
(i)The options on the selection box are named according to the following rules:
  • ML algorithm - hyperparameter number - hyperparameter name
  • eg: SVM-3-[C,kernel,gamma]
(ii)If you click the Advanced button, a settings panel will show up. You can set hypetparameter range by inputting
Minimum and Maxmum in the Min Input box and Max input box respectively.Make sure you input the reasonable range of the hyperparameter, otherwise it may cause errors .If you are not familiar with the hyperparameter, just skip this step because reasonable ranges of hyperparameters have already been preset.
⑧Optimization times (optional): (default: 100).
Input a number to decide the optimization times of hyperparameters.
⑨Make sure you have followed the guide and submit the job.

Online prediction

①Input a string with letters, numbers or underline as your unique jobname.
②Upload a ZIP file for the calculation of descriptors:
The file requirements are exactly the same as the above-mentioned Descriptors Generation module.
③Choose a target:
To use the well-constructed SFs, users should choose the corresponding model.
But for the prediction using the generic model, do not need to choose any model.
④Make sure you have followed the guide and submit the job.

Results

Once users submitted their job, they will get a pdf file (result_url.pdf) immediately, which is roughly divided into the following two parts:
①Information:
The basic information of the users’ input.
②Urls:
Two steps to get the results:
  • Check jobs’ status until the status shows ‘Completed’.
  • Click the Visualization pages’ url or Result_download page’s url to get the results.

Queue

①name: your jobname
②type: function type
  • Descriptors: descriptors generation
  • AI-scoring-function: AI-based scoring function construction
  • Prediction: online prediction
③submitted time: the system time when you submitted your job
④status:
  • Queue: the job has been submitted and wait in queue for calculation.
  • Running: the job is on calculation.
  • Completed: the job has finished and users can view or download the results.
  • Error: Some errors happened during the calculation.

Visualization

①Res_label: If this checkbox is selected, the label of the protein residues will be added to the Visualization box.
(This may cause a freezing of the web pages.)
②Surface: Select this checkbox to add a surface of the protein and the ligand.
③Complex: If the checkbox of this column is checked, the corresponding 3D model of the protein-ligand complex will
show up in the Visualization box.
④Ligand: If the checkbox of this column is checked, the corresponding 3D model of the ligand wiil show up in the
Visualization box.
⑤Visualization: The 3D model of the ligand or the protein-ligand complex will be shown here.
For the results of AI scoring function construction and online prediction, the ligand cheboxes will be shown in two tables,
named Inhibitors and Non-inhibitors, respectively.
For the results of a target-specific job, you can select multiple checkboxes at the same time to compare the differences of
their binding modes.

Contact

If you have any questions, you can get our contacting information or you can just send email on the contact page by simply clicking the button named Send Message after inputting your text.

Documentation

If you want to know more about ASFP, you can click here to download the documentation.