metadata
license: mit
tags:
- chemistry
- biology
pretty_name: PeptideMTR Pretraining Data
size_categories:
- 100M<n<1B
PeptideMTR Training Data
This repository contains the dataset for the PeptideMTR paper. It is designed for SMILES encoder models trained by masked-language modeling (MLM) and/or multi-target regression (MTR) tasks, focusing on mapping peptide sequences to biochemical properties.
Link to the manuscript will be added here when available.
Dataset Summary
The dataset includes peptide sequences paired with 99 RDKit-derived descriptors representing various physicochemical properties (e.g., molecular weight, LogP, surface area, and charge descriptors).
Data Structure
SMILES: The SMILES representation of the molecule.descriptors: 99 continuous numerical features generated via RDKit.
Usage
To use this dataset with the Hugging Face datasets library:
from datasets import load_dataset
ds = load_dataset("your-username/PeptideMTR_training_data")