Comments (40)
I recommend you to assess our method on your own target in advance, e,g., docking power (redocking, cross-docking), screening power (testing whether our approach could distingush active compounds from inactive componds for your own target).
from rtmscore.
hi,I would like to ask you a question, Can I use it to screen large molecular databases?
Thanks!
from rtmscore.
Of course you can have a try.
from rtmscore.
The database given in the literature cannot be downloaded. Is there any way to solve it?
thanks!
from rtmscore.
I can access all the links provided in the litertature (http://www.pdbbind.org.cn/; https://
zenodo.org/record/5525936; https://zenodo.org/record/6859325). Perhaps there is something wrong with your network.
from rtmscore.
Ok, thank you for your answer.
from rtmscore.
from rtmscore.
You can directly use the code provided as the example for rescoring, without any changes. Just see our method as a generic scoring function like X-Score or AutoDock Vina. However, it should be noticed that our model in this version just serves as a scoring function rather than a docking program, so you need to generate the binding poses first. Additionally, it is better for you to generate multiple poses (e.g., 10) for a certain ligand for our model to select the most reasonable pose for rescoring, which may achieve better performance.
from rtmscore.
from rtmscore.
Yes, you need to generate the binding poses between the each ligand in the database and a protein.
Of course you can, but it is better for you to assess our model on your own target first, since the performance of our method may be highly target-specific.
from rtmscore.
from rtmscore.
As shown in Figure 4 in our paper, the virtual screening performance of our method may vary a lot when different targets are focused on. Hence of you try our method on your own target, I'm unclear how will it actually perform and thus I suggest you to conduct an initial assessment first.
from rtmscore.
from rtmscore.
from rtmscore.
It is defined as the percentage of true binders observed among all of the true binders for 1% of the top-ranked candidates of
a chemical library.
The bigger means the better.
Not necessarily. It is just a metric for retrospective study, but not necessarily suitable for your own system in real-world applications since both the size of the screening library and the number of compounds you tend to select for purchase/synthesis are unknown. The simplest method to judge the performance of our model for your target may just see whether the model can make the active compounds lie in the front rank.
from rtmscore.
from rtmscore.
This file is downloaded from http://www.pdbbind.org.cn/casf.php.
Actually I cannot understand what you mean to do.
from rtmscore.
from rtmscore.
from rtmscore.
You can directly use our trained models to predict the binding poses. But first, you should use the external docking program (e.g., AutoDock Vina, Glide, GOLD, etc.) to generate multiple docking poses for your own ligands and the target, and then use our method for rescoring, just as described in our demo files.
from rtmscore.
from rtmscore.
You can generate the pocket using the "extract_pocket" function in RTMScore/feats/extract_pocket_prody.py.
from rtmscore.
from rtmscore.
You should call the "extract_pocket" function in RTMScore/feats/extract_pocket_prody.py and write a script by yourself rather than directly run the script.
You can use some multiprocessing utilities (e.g., multiprocessing module in Python) to handle the molecules simultaneously.
from rtmscore.
from rtmscore.
It is calculated as:
You can refer to the following script:
11.zip
from rtmscore.
from rtmscore.
You can refer to the following script.
generate_pocket.zip
from rtmscore.
from rtmscore.
You can just use the rtmscore.py for prediction. Just do as the demo.
from rtmscore.
from rtmscore.
The score represents the binding strength of a certain ligand conformation to the target, and the higher the better. Multiple docking poses of a certain molecule should be generated first, and then rescored by rtmscore.py. The best-scored pose of a certain molecule can be selected as the final binding pose of that molecule, and this score could represent the binding of the molecule to the target.
If you use our method to screen the database, you can first enrich the database to a certain number of top-ranked molecules (e.g., 500), and then determine your compounds from these molecules just based on your experience.
from rtmscore.
from rtmscore.
" Can't the model directly screen or rank based on the score of the binding pose?" Of course you can. But I think you can never completely rely on the docking scores when conducting a virtual screening campaign.
from rtmscore.
from rtmscore.
from rtmscore.
from rtmscore.
1qkt_p.pdb the structure of the protein
1qkt_l.sdf the pose of crystal ligand used to define the pocket
1qkt_decoys.sdf the poses of the molecules that you want to score
from rtmscore.
from rtmscore.
from rtmscore.
Related Issues (19)
- error of create conda env HOT 3
- Meaning of the distance threshold HOT 2
- casf对接和筛选能力 HOT 10
- Pretrained models - question HOT 2
- rdkit cannot load some mol2 files HOT 7
- 环境问题 HOT 6
- RTMScore的score的打分如何查看?
- 使用新的口袋和配体测试Score为0 HOT 1
- re-train a model based on new dataset
- The AUC values presented in Table 6 and Figure 6A seem to be inconsistent. HOT 2
- How to Batch Process RTMScore for Multiple Models in PDB and SDF Files
- Could you provide raw data of pdbbind2020? HOT 1
- integrating w/ docking HOT 1
- data preprocess script for training on new dataset HOT 7
- Some of the ligands in CASF core set cannot be read by RDKit successfully HOT 2
- Additive terms in MDN outputs HOT 2
- How should we preprocess .pdbqt files before using RTMScore? HOT 6
- Graph files on zenodo HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from rtmscore.