ncats / ncats-adme Goto Github PK

The source code for ADME@NCATS application that hosts prediction models for ADME properties. Link to application: https://opendata.ncats.nih.gov/adme/home

JavaScript 0.26% TypeScript 6.78% HTML 43.44% SCSS 0.90% Python 35.43% Dockerfile 0.09% CSS 12.94% Makefile 0.06% Batchfile 0.08% Shell 0.01%

ncats-dpi-ifx adme machine-learning artificial-intelligence deep-learning

ncats-adme's Introduction

ADME@NCATS

This repository contains the code for the application ADME@NCATS which hosts QSAR models for different ADME endpoints. To use the application locally, you can either use Git to clone the respository, or you can simply download a ZIP file (by clicking the green "Code" button on the top right corner) and then unzip it. The next steps are described below.

If you use Git to clone this repository, please use the --recursive flag:

git clone --recursive https://github.com/ncats/ncats-adme.git

If you download the application, you also need to download and unzip chemprop, but make sure to unzip the contents of chemprop inside the server folder so the the ncats-adme folder/file structure looks something like this:

ncats-adme
- client
- server
  - chemprop
    - chemprop
    - docs
    - scripts
    - ...

Models will be loaded from NCATS servers so you will need access to the internet when you first run the application. Alternatively, if you want to download the models, the files are available as follows:

Installing required software

Install anaconda or miniconda

Python is also required but it is included with either installation of conda or miniconda.

Setting up the environment

You only have complete these steps one time.

Open your terminal

If you're on Windows, open Anaconda Prompt (window -> Anaconda3 -> Anaconda Prompt)
If you're on Mac or Linux, open your Terminal

Change the working directory (windows or Mac and Linux) to where you have ncats-adme and then go (CD one more time) into the server directory
Create environment

For Windows and Linux machines
1. Type conda env create --prefix ./env -f environment.yml and hit Enter
2. Wait several minutes for the envitonment to be created
3. For Windows machines only, type pip install typed-argument-parser and hit Enter
For Mac machines
1. Type conda env create --prefix ./env -f environment_mac.yml and hit Enter
2. Wait several minutes for the envitonment to be created

Running the application

If you're doing this immediately after completing the steps above, skip to step 4
Open your terminal

If you're on Windows, open Anaconda Prompt (window -> Anaconda3 -> Anaconda Prompt)
If you're on Mac or Linux, open your terminal

Change the working directory (windows or Mac and Linux) to where you have ADME_RLM and then go into the server directory
Type conda activate ./env and hit Enter
Type python app.py and hit Enter
Open Chrome or Firefox and browse to http://127.0.0.1:5000/
To close the application, hit Ctrl + c or Cmd + c in the Terminal and then type conda deactivate and hit Enter to close the conda environment

ncats-adme's People

Contributors

Stargazers

Watchers

Forkers

sailfish009 aojesanmi mapleknight kannabhargav carrystar rnaimehaom dhanshreea iwwwish emmakodes inyrkz ajoke23 akadapa chevvak2 pabbathreddya2

ncats-adme's Issues

Footer position

The footer does not stick to the bottom on some of the pages of the website. Example:

SMILES input

Implement SMILES input (e.g. a text field) so that users can directly provide a SMILES string as input instead of obtaining a MOL file for a molecule of interest.

Error: I'M A TEAPOT

For some of the molecules i am getting Error: I'M A TEAPOT . How to resolve this issues or understand why/what casuse this error?

<!doctype html>
<html lang=en>
<title>418 I&#39;m a teapot</title>
<h1>I&#39;m a teapot</h1>
<p>There was an unknown error.</p>

Option to run it locally as command line without browser

Hi,

I am interested in using the tool as a command line. Wondering if this feature will/can be made available?

OR any suggestion to running as a command line using the current version of the tools?

Regards,
Ravin

Prediction probabilities need explanation

A recent paper reported incorrect AUC value for our RLM model due to the ambiguity associated with the prediction probabilities. Probability definitions need to be clearly explained on the website.

Conda env cannot be created with the provided yml file

I am on the mac (M1), and getting following error. Any suggestion on how to fix the issues?

Collecting package metadata (repodata.json): | WARNING conda.models.version:get_matcher(535): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.8.0.*, but conda is ignoring the .* and treating it as 1.8.0
WARNING conda.models.version:get_matcher(535): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.9.0.*, but conda is ignoring the .* and treating it as 1.9.0
done
Solving environment: failed

ResolvePackageNotFound:
  - pytorch=1.6.0
  - rdkit=2020.03
  - torchvision=0.7.0
  - scikit-learn=0.22
  - numpy=1.18
  - efelix::fpsim2=0.2.6
  - matplotlib=3.3.0```

Model download links

Hi! Thanks a lot for your resource!
The download links seem broken (again?) though.

One button click to download predictions from all models

Feature requested by a user via email (see below).

Thank you. That worked.

Also, can a single table, including all calculated properties, be downloaded?

Thanks,
***

From: ***
Sent: Friday, September 9, 2022 11:40 AM
To: *******; ADME-NCATS <[[email protected]](mailto:[email protected])>
Subject: Re: Inputting SMILES?

Hi ***,

If you have a SMILES string, you should be able to directly paste it into the molecule editor. Make sure to click once within the editor region and then paste using the keyboard shortcut (Ctrl + V or Cmd + V).

Please let me know if this does not solve the issue.

Best regards,
Vishal

From: ***
Date: Friday, September 9, 2022 at 11:37 AM
To: ADME-NCATS <[[email protected]](mailto:[email protected])>
Subject: Inputting SMILES?

How do you input a SMILES string into the text box?

no information about redkit dependencies in ReadMe

Hello maintainers! I am Medhavi Singh , an Outreachy applicant. After setting up the environment, when I tried running "python app.py"
I received following error.

I think it should be mentioned in the ReadMe that we need to install redkit before proceeding. This can be updated. Do let me know your thoughts. If you want I can create a PR for updated ReadMe.
Thank you!

Implement SocketIO

Implement web sockets on flask and angular in order to avoid server timeouts.

Flask:
https://flask-socketio.readthedocs.io/en/latest/intro.html

Angular:
https://www.npmjs.com/package/ngx-socket-io

can ncats-adme be run locally in commandline without calling API?

Hi,
I am wondering if there is an option to run the ncats-adme model locally without calling for API. I have dockerized the tool, and exposed API to the port locally, then once api becomes available, I am running the model. IF there is way to run the model without going via the API, i think it will be much effiecient, computationally.

Regards,
Ravin

Cannot make cyp450 prediction

When I try to make cyp450 prediction. It went collapse on my own computer. I followed all the steps carefully and nothing went wrong. But still cannot precit cyp450 property. What should I do?

Failure to run the app.py file

Hi @iwwwish, I'm encountering an error while trying to run the application.

Error encountered while executing the python app.py command

Whenever I try to run python app.py, I get _pickle.UnpicklingError: invalid load key, '<'. error

Operating System

WSL on Windows 10
Python 3.8.13
Anaconda 4.14.0

Steps I took

Installed Anaconda
Cloned the repo git clone --recursive https://github.com/ncats/ncats-adme.git
cd ADME-NCATS
Created environment conda env create --prefix ./env -f environment.yml
Executed the pip install typed-argument-parser command
Activated environment conda activate ./env
Got the above error when running command python app.py

Model download links broken

The downloading links provided in README seem broken. They were working ok about a week ago.

Output file contains more rows than input file

When the input file contains duplicates (SMILES or IDs) the prediction result file returned by the application contains more rows than input.

LICENSE file

Sorry, one more request: Could you add a LICENSE file to the repository?

Keras dependency

Get rid of keras dependency since we do not host DNN and LSTM models anymore.

setting up local version

Apologies - chemist with a little computer knowledge rather than the other way around. I have been attempting to set up local version, however when running app.py as described I get an error. As far as I can tell from init.py the script is attempting to download model files that are no longer on tripod.nih.gov ( eg rf_morgan_all_data.pkl from https://tripod.nih.gov/pub/adme/models/rlm/ ). Is this the case, if so is there a workaround?

Cannot parse SMILES strings containing positive charge

Example: Niclosamide
Input SMILES: OC1=C(C=C(Cl)C=C1)C(=O)NC2=CC=C(C=C2Cl)[N+]([O-])=O
Error: SMILES Parse Error: syntax error while parsing: C1=CC(O)=C(C(=O)NC2=C(Cl)C=C([N

ModuleNotFoundError: No module named 'predictors.rlm'

Hi there,

I'm new to this repo, I followed the installation instructions from the readme and then I tried to run the command python app.py in the server folder but I got the following error:

(/home/aml/ncats-adme/server/env) aml@aml-HP-ZBook-15-G3:~/ncats-adme/server$ python app.py
Traceback (most recent call last):
  File "app.py", line 20, in <module>
    from predictors.rlm.rlm_predictor import RLMPredictior
ModuleNotFoundError: No module named 'predictors.rlm'

Can you please help me with that?

HLM Model File

Hello @iwwwish !

Thanks for sharing this resource with the community, it is a great implementation! I am the CEO of Ersilia a non profit organisation working to support infectious disease research in LMIC.
I have successfully installed all models and run them in my system, but the HLM model seems to be corrupt (I am downloading the files from the link in the development branch)
The model is not either available for online inference through the ADME-NCATS website.
Could you check the current model and let me know if there is an updated version?

Many thanks

ResolvePackageNotFound: - gunicorn

On windows system, when we try to resolve environment using conda, following error shows up

ResolvePackageNotFound:

gunicorn