Frequently Asked Questions

Disclaimer: iDrug platform respect customer's private information and data, and we will also try our best to protect data security that customer provided in our platform. Moreover, we will delete such data information upon customers request.

1. Can I use iDrug ADMET prediction module to predict peptides, proteins, and other macromolecules?
2. How many molecules can I submit to iDrug platform?
3. Should I name the inputted molecules?
4. Can I draw more than one molecular structure in the molecule editor at a time?
5. Does aromatic or kekule representation change anything for prediction?
6. Is it preferable to input the neutral form of the molecule?
7. Can I cut-and-paste a molecule for input?
8. Can I input a SMARTS string to do ADMET prediction?
9. How long will it take for computing 50 molecules?
10. How are computed molecular descriptors and physicochemical properties?
11. Can one export the output panel in a PowerPoint slide or as a picture?
12. What are the differences between step 1 and 2 in iDrug Bioactivity prediction module?
13. How many targets does iDrug Bioactivity prediction module support to do virtual screening?
14. How many assays related to safety panel and kinase selectivity panel does iDrug Bioactivity prediction module contain?
15. Can I choose more than one compound libraries at one time for virtual screening?
16. Are your compound libraries in Bioactivity prediction module up-to-date?
17. How many molecules can I export in the virtual screening module?
18. Can no reference compound be uploaded in generative chemsitry module?
19. How many molecules can be generated at one time for generative chemistry?
20. How the algorithm keeps the possible actitivity of the generated molecules?
21. What are the difference between 2D and 3D similarities?
22. Can customers upload more than one reference compound in generative chemsitry?
23. Can generative chemistry and virtual screening support customers training the model by their own data in iDrug platform?
24. What is the data set for training the retrosynthesis model?
25. Can iDrug's retrosynthesis module predict complex natural products?
26. How long will it take for predicting the synthetic route of one molecule?
27. How to determine the terminatal of retrosynthesis prediction?

1. Can I use iDrug ADMET prediction module to predict peptides, proteins, and other macromolecules?

It is feasible in practice if the structure can be described as a SMILES notation. However, most of the proposed models have an applicability domain close to druglike organic compounds with very short oligopeptides or short oligosaccharides. For severely different molecular structures, it is unlikely that the information returned by the predictive methods would be accurate. iDrug ADMET module is intended to be employed in the context of drug discovery and medicinal chemistry. Any other usage requires extreme caution in the interpretation of data.

2. How many molecules can I submit to iDrug platform?

For ADMET prediction module, you can input one molecule by drawing the structure or input one or two smiles of the molecules. You can also upload a file containing molecular smiles. Do NOT exceed 100 entries per list. You can divide your list and calculate for several times if you have more than 100 molecules. Do NOT launch several calculations at the same time, please wait for the whole calculation to be terminated before running the next batch. You are welcome to run such batch calculations one after the other if the total number of molecules submitted does not exceed 1,000. For Bioactivity prediction module, as it requires long computing resources, we currently support you to submit not more than 10 screening jobs for one account a week. You can contact us to find collaborations if you have many virtual screening projects.

3. Should I name the inputted molecules?

It is not mandatory. You need to make sure there is “smiles” or “SMILES” or “Smiles” title in your table. However it is recommended for the sake of ease in the output analysis since the name of the molecule appears as a title in the output panel and as a legend in the Graphical output.

4. Can I draw more than one molecular structure in the molecule editor at a time?

You should not draw multiple unconnected molecular fragments in the editor because they would be considered as a unique structure and appear in one single row separated by a dot in the SMILES list. A correct input list must involve two different molecules separated by a comma’,’.

5. Does aromatic or kekule representation change anything for prediction?

No. Either aromatic or kekule description can be inputted without impacting the output values as one first step in the process is a standardization of the molecular structure including dearomatization (i.e. kekulization).

6. Is it preferable to input the neutral form of the molecule?

Yes, definitively. The SMILES entry is taken as given and not neutralized explicitly for every model. As many of the proposed models are trained on the neutral form of compounds, submitting an ionized structure would lead to severe biases in the predictions unless it is a permanent ion or a zwitterion.

7. Can I cut-and-paste a molecule for input?

You can easily paste SMILES in the input list text field.

8. Can I input a SMARTS string to do ADMET prediction?

Current, we only accept SMILES string as an input.

9. How long will it take for computing 100 molecules in ADMET modules?

The time depends on the models we included. The more models we have, the longer time it will take. Currently, we have launched 50 models (12 basic properties and 38 ADMET models). For 100 molecules, it usually takes about 3 minutes.

10. How are computed molecular descriptors and physicochemical properties?

IDrug ADMET prediction relies on the definition of RDKit 2019.03.4 for molecular descriptors like atom/bond/group counts, which is an open-source cheminformatics software.

11. Can one export the output panel in a PowerPoint slide or as a picture?

Not directly at the moment. However the panels are made to fit well a standard computer display. So a screenshot can be pasted in any presentation application then. The default format for results output is csv file.

12. What are the differences between step 1 and 2 in iDrug Bioactivity prediction module?

In step 1, you need to choose related assays to your target of interest. You can think it as choose ‘on target’. In step 2, we support you to check safety related assays, in which you can check some potential off-target toxicities of the molecules.

13. How many targets does iDrug Bioactivity prediction module support to do virtual screening?

Currently, we support 920 targets and 2224 assays to do virtual screening on.

14. How many assays related to safety panel and kinase selectivity panel does iDrug Bioactivity prediction module contain?

We included 34 safety related assays and 99 kinase selectivity assays for off-target check.

15. Can I choose more than one compound libraries at one time for virtual screening?

Yes, you absolutely can. We support you to select commercial libraries. We also did pre-calculations on the databases we supported for the sake of saving time. However, more compound libraries you choose will take a longer time for computing. Be patient!

16. Are your compound libraries in Bioactivity prediction module up-to-date?

Yes, we will update the compound libraries regularly, but may not be very often. The last update is Mar 1st, 2020.

17. How many molecules can I export in the virtual screening module?

Considering the time for downloading the data, we suggest you to control your final selected compounds no more than 300,000.

18. Can no reference compound be uploaded in generative chemsitry module?

Of cource you can. You can choose not to upload the reference compound. You can only choose to select the target that you are intersted in. Our algorithm will refine the structure of generated molecules by reference compounds.

19. How many molecules can be generated at one time for generative chemistry?

Considering the time for generative chemistry process, our algorithm currently will generate 1000-2000 molecules. The number will also depends on the target and reference compounds that customer provided.

20. How the algorithm keeps the possible actitivity of the generated molecules?

Our algorithm generates molecules by sampling at the gaussian distribution of known active compounds related to the targets. Therefore, the generated molecules theoretically will keep the potential activity. But we still stongly suggest you do further activity scoring by other models, for example, our LBDD module.

21. What are the difference between 2D and 3D similarities?

2D similarity is calulated based on Morgen fingerprint, while 3D similarity is calculating the shapes of molecules and comparing with reference compounds. You can refer to the literature if you want to know more about 3D similarity: J. Chem. Inf. Model. 2020, 60, 1983−1995

22. Can customers upload more than one reference compound in generative chemsitry?

Currently, we do not support uploading more than one reference compound. We will reliaze that in our updated version.

23. Can generative chemistry and virtual screening support customers training the model by their own data in iDrug platform?

We currently do not support customer training their own model. But we will support you in very near feature next year. You can contact us for collaboration for such purpose.

24. What is the data set for training the retrosynthesis model?

The data set used for training the model is open-source USPTO 480K data set.

25. Can iDrug's retrosynthesis module predict complex natural products?

The current beta version of retrosynthesis module is limited by the data set. The data is all from patent data. Therefore, we highly suggest for medicinal chemistry. Predictions for complex natural products may not be accurate enough.

26. How long will it take for predicting the synthetic route of one molecule?

It usually will take 3 to 4 minutes for one molecule.

27. How to determine the terminatal of retrosynthesis prediction?

Over 20 millions molecules from commercial chemicals or building blocks are collected for the terminating the prediction. You can also continue with predition of complex terminal reactants.