Google Colab Integration for Protein Design -- 2 (Max 70 USD)
$30-250 USD
Plačilo ob prevzemu
I'm looking to make a protein design pipeline accessible via Google Colab, primarily enabling non-experts to utilize it. The goal is to democratize access to this tool. The method was published recently by a group in the US in Nature Machine Intelligence [login to view URL], with code and model weights available under MIT license:
[login to view URL]
We are especially interested in their "python [login to view URL]" scripts, and would like to make them as interactive as other protein design scripts like: [login to view URL]
Here simple tabular selection, input boxes and such make it very user-friendly.
The objectives of the project:
1) Working inference scripts and fast loading of model + dependency setup.
2) Uploading target .pdb and .sdf files
3) Adjustable slides and tabular selection for hyper parameters (N_samples, temperature, etc)
4) Writing an analysis dataframe --> .csv with binding energies and PLIP interaction analysis.
5) Zip and Download results.
Key Project Requirements:
- Modifying the protein design pipeline for Google Colab: This involves ensuring the pipeline is fully functional on the platform and user-friendly for non-experts.
-Implement the necessary features from above.
User Input Requirements:
- The pipeline will need to accept .pdb and .sdf files, which will parameterize the protein and ligand pose.
Ideal Skills & Experience:
- Proficiency in Python and Google Colab and dependency wrangling.
- Some experience with protein design pipelines, like AF2 are a benefit.
- Ability to simplify complex processes for non-expert users, with minimal user interface design for simplicity (using available Google Colab functionality)
P.S. The model weights can be grabbed directly using gdown using the file ID provided in the GitHub:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
import os
import gdown
# Ensure the 'checkpoint' directory exists
[login to view URL]('./checkpoint', exist_ok=True)
# Google Drive file ID
file_id = "PUTFILEIDHERE"
# Construct the download URL
url = f"[login to view URL]{file_id}"
# Filepath to save the downloaded file
output = './checkpoint/downloaded_file'
# Download the file
[login to view URL](url, output, quiet=False)
print(f"File downloaded to {output}")
ID projekta: #38881051
Več o projektu
18 freelancerjev ponuja v povprečju za $148 na tem delu
Dear Hiring Manager, I am Abdul, a seasoned Python developer with extensive experience in software testing and debugging. I have a strong background in developing and implementing various Python projects, including th Več
Hello I have read out the details of your project. And I am one of the suitable candidates for your project. We have more than 9 years exp in the development. We have focused on delivering perfect valuable delivera Več
Hello Dear! Good Day! Hope you are doing fine. This is Toriqul Islam . I am an expert "Web Developer" with 10+ years of working experience in PHP, HTML5, CSS3, JavaScript, jQuery, Bootstrap, MySql and different Frame Več
Hello, Drawing from my technical fluency with Python and Google Colab, I firmly believe I'm the ideal fit to make your protein design pipeline accessible and user-friendly on Google Colab. My experience with back-end Več
Hello, Niklas M. ? This project"Google Colab Integration for Protein Design -- 2 (Max 70 USD)" seems like a very good fit for me. I think the loud introduction is pretentious. Actions are more important than words. I Več
❤️❤️Hi Niklas M.,❤️❤️ I am Artur, a seasoned Web Developer with nearly 6 years of experience in Python, Debugging and Software Testing. After reviewing your project details, I am confident in my ability to design and Več
Hello, As a seasoned Full Stack Developer with specialized knowledge in Python, I'm confident in my ability to effectively support and enhance the protein design pipeline you're aiming to integrate with Google Colab. Več
Hi. I am an AI scientist and also a Mathematician. I have strong foundation in linear algebra, calculus, probability, and statistics, which are essential for understanding DL algorithms and model performance. I think Več
