BindNet

BindNet is a deep learning-based protein–ligand interaction analysis tool designed to help researchers rapidly identify potential drug molecules from protein-observed NMR (HSQC) experiments, accelerating drug screening and early-stage drug discovery. Users simply upload their spectra, and the model automatically performs the analysis and generates prediction results.

Usage

Step 1. Export HSQC spectra

In TopSpin, open the following spectra separately:

• Protein HSQC spectrum (Protein)
• Corresponding Protein + Ligand HSQC spectrum (Protein + Ligand)

For each spectrum, select:

Save data of currently displayed region in a text file

to export the currently displayed region as a .txt file.

Step 2. Organize the data

Place each matching pair of spectra into the same folder:

• One Protein spectrum
• One corresponding Protein + Ligand spectrum

Repeat this process for all samples.

Step 3. Compress and upload

Place all sample folders into a single parent folder and compress the parent folder into a ZIP file.

Upload the ZIP file to the website, click Run, and wait for the model to complete the analysis and generate the prediction results.

File Naming Requirements (Important)

To ensure the output CSV is generated correctly and remains well organized, please follow the naming rules below.

• Each sample folder must be named using only numbers (e.g., 1, 2, 3, …).
• Each TXT filename must end with _number.
• The TXT file with the smaller ending number represents the Protein spectrum.
• The TXT file with the larger ending number represents the Protein + Ligand spectrum.
• Within each sample folder, the ending number of the Protein spectrum must be smaller than that of the corresponding Protein + Ligand spectrum.

Example

test_data.zip
└── test_data
    ├── 1
    │   ├── ProteinA_2.txt
    │   └── ProteinA_LigandA_concentration_100_4.txt
    │
    ├── 2
    │   ├── ProteinA_2.txt
    │   └── ProteinA_LigandA_concentration_200_6.txt
    │
    └── 3
        ├── ProteinB_2.txt
        └── ProteinB_LigandB_concentration_100_4.txt

In this example:

• 1, 2, and 3 are sample folders named using only numbers.
• ProteinA_2.txt is the Protein spectrum and the filename ends with _2.
• ProteinA_LigandA_concentration_100_4.txt is the corresponding Protein + Ligand spectrum and the filename ends with _4, and 4 > 2.

Output CSV Columns