Skip to content

Commit a9d639d

Browse files
committed
Update to version 1.1.0 of DeepInteract and update Zenodo links
1 parent d0781de commit a9d639d

File tree

5 files changed

+37
-27
lines changed

5 files changed

+37
-27
lines changed

Diff for: README.md

+33-23
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22

33
# DeepInteract (ICLR 2022)
44

5-
[![Paper](http://img.shields.io/badge/paper-arxiv.2110.02423-B31B1B.svg)](https://openreview.net/forum?id=CS4463zx6Hi) [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.5797024.svg)](https://doi.org/10.5281/zenodo.5797024)
5+
[![Paper](http://img.shields.io/badge/paper-arxiv.2110.02423-B31B1B.svg)](https://openreview.net/forum?id=CS4463zx6Hi) [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.6299835.svg)](https://doi.org/10.5281/zenodo.6299835)
66

77
[<img src="https://twixes.gallerycdn.vsassets.io/extensions/twixes/pypi-assistant/1.0.3/1589834023190/Microsoft.VisualStudio.Services.Icons.Default" width="50"/>](https://pypi.org/project/DeepInteract/)
88

@@ -248,8 +248,8 @@ Now that we know Docker is functioning properly, we can begin building our Docke
248248
249249
```bash
250250
mkdir -p project/checkpoints
251-
wget -P project/checkpoints https://zenodo.org/record/5797024/files/LitGINI-GeoTran-DilResNet.ckpt
252-
wget -P project/checkpoints https://zenodo.org/record/5797024/files/LitGINI-GeoTran-DilResNet-DB5-Fine-Tuned.ckpt
251+
wget -P project/checkpoints https://zenodo.org/record/6299835/files/LitGINI-GeoTran-DilResNet.ckpt
252+
wget -P project/checkpoints https://zenodo.org/record/6299835/files/LitGINI-GeoTran-DilResNet-DB5-Fine-Tuned.ckpt
253253
```
254254
255255
3. Build the Docker image (Warning: Requires ~13GB of Space):
@@ -400,18 +400,39 @@ referenced in `project/datasets/builder/psaia_config_file_input.txt`.**
400400
401401
### Download training and cross-validation DGLGraphs
402402
403-
To train, retrain, or cross-validate DeepInteract models using DIPS-Plus and/or CASP-CAPRI targets, we first need to download the preprocessed DGLGraphs from Zenodo:
403+
To train, fine-tune, or test DeepInteract models using CASP-CAPRI, DB5-Plus, or DIPS-Plus targets, we first need to download the preprocessed DGLGraphs from Zenodo:
404404
405405
```bash
406-
# Download and extract preprocessed DGLGraphs for DIPS-Plus and CASP-CAPRI
406+
# Download and extract preprocessed DGLGraphs for CASP-CAPRI, DB5-Plus, and DIPS-Plus
407407
# Requires ~55GB of free space
408-
mkdir -p project/datasets/DIPS/final
409-
cd project/datasets/DIPS/final
408+
# Download CASP-CAPRI
409+
mkdir -p ../../CASP_CAPRI/final
410+
cd ../../CASP_CAPRI/final
411+
wget https://zenodo.org/record/6299835/files/final_raw_casp_capri.tar.gz
412+
wget https://zenodo.org/record/6299835/files/final_processed_casp_capri.tar.gz
413+
414+
# Extract CASP-CAPRI
415+
tar -xzf final_raw_casp_capri.tar.gz
416+
tar -xzf final_processed_casp_capri.tar.gz
417+
rm final_raw_casp_capri.tar.gz final_processed_casp_capri.tar.gz
418+
419+
# Download DB5-Plus
420+
mkdir -p ../../DB5/final
421+
cd ../../DB5/final
422+
wget https://zenodo.org/record/6299835/files/final_raw_db5.tar.gz
423+
wget https://zenodo.org/record/6299835/files/final_processed_db5.tar.gz
424+
425+
# Extract DB5-Plus
426+
tar -xzf final_raw_db5.tar.gz
427+
tar -xzf final_processed_db5.tar.gz
428+
rm final_raw_db5.tar.gz final_processed_db5.tar.gz
410429
411430
# Download DIPS-Plus
412-
wget https://zenodo.org/record/5797024/files/final_raw_dips.tar.gz
413-
wget https://zenodo.org/record/5797024/files/final_processed_dips.tar.gz.partaa
414-
wget https://zenodo.org/record/5797024/files/final_processed_dips.tar.gz.partab
431+
mkdir -p project/datasets/DIPS/final
432+
cd project/datasets/DIPS/final
433+
wget https://zenodo.org/record/6299835/files/final_raw_dips.tar.gz
434+
wget https://zenodo.org/record/6299835/files/final_processed_dips.tar.gz.partaa
435+
wget https://zenodo.org/record/6299835/files/final_processed_dips.tar.gz.partab
415436
416437
# First, reassemble all processed DGLGraphs
417438
# We split the (tar.gz) archive into two separate parts with
@@ -423,17 +444,6 @@ cat final_processed_dips.tar.gz.parta* >final_processed_dips.tar.gz
423444
tar -xzf final_raw_dips.tar.gz
424445
tar -xzf final_processed_dips.tar.gz
425446
rm final_processed_dips.tar.gz.parta* final_raw_dips.tar.gz final_processed_dips.tar.gz
426-
427-
# Download CASP-CAPRI
428-
mkdir -p ../../CASP_CAPRI/final
429-
cd ../../CASP_CAPRI/final
430-
wget https://zenodo.org/record/5797024/files/final_raw_casp_capri.tar.gz
431-
wget https://zenodo.org/record/5797024/files/final_processed_casp_capri.tar.gz
432-
433-
# Extract CASP-CAPRI
434-
tar -xzf final_raw_casp_capri.tar.gz
435-
tar -xzf final_processed_casp_capri.tar.gz
436-
rm final_raw_casp_capri.tar.gz final_processed_casp_capri.tar.gz
437447
```
438448
439449
Navigate to the project directory and run the training script with the parameters desired:
@@ -455,8 +465,8 @@ cd "$DI_DIR"
455465
456466
# Download our trained model checkpoints
457467
mkdir -p project/checkpoints
458-
wget -P project/checkpoints https://zenodo.org/record/5797024/files/LitGINI-GeoTran-DilResNet.ckpt
459-
wget -P project/checkpoints https://zenodo.org/record/5797024/files/LitGINI-GeoTran-DilResNet-DB5-Fine-Tuned.ckpt
468+
wget -P project/checkpoints https://zenodo.org/record/6299835/files/LitGINI-GeoTran-DilResNet.ckpt
469+
wget -P project/checkpoints https://zenodo.org/record/6299835/files/LitGINI-GeoTran-DilResNet-DB5-Fine-Tuned.ckpt
460470
```
461471
462472
### Predict interface contact probability maps

Diff for: project/datasets/CASP_CAPRI/casp_capri_dgl_dataset.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -251,4 +251,4 @@ def raw_path(self) -> str:
251251
@property
252252
def url(self) -> str:
253253
"""URL with which to download TAR archive of preprocessed pairs."""
254-
return 'https://zenodo.org/record/5546775/files/final_processed_casp_capri.tar.gz'
254+
return 'https://zenodo.org/record/6299835/files/final_processed_casp_capri.tar.gz'

Diff for: project/datasets/DB5/db5_dgl_dataset.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -254,4 +254,4 @@ def raw_path(self) -> str:
254254
@property
255255
def url(self) -> str:
256256
"""URL with which to download TAR archive of preprocessed pairs."""
257-
return 'https://zenodo.org/record/5546775/files/final_raw_db5.tar.gz?download=1'
257+
return 'https://zenodo.org/record/6299835/files/final_processed_db5.tar.gz'

Diff for: project/datasets/DIPS/dips_dgl_dataset.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -276,4 +276,4 @@ def raw_path(self) -> str:
276276
@property
277277
def url(self) -> str:
278278
"""URL with which to download TAR archive of preprocessed pairs (Need to manually download Part B)."""
279-
return 'https://zenodo.org/record/5546775/files/final_processed_dips.tar.gz.partaa'
279+
return 'https://zenodo.org/record/6299835/files/final_processed_dips.tar.gz.partaa'

Diff for: setup.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@
44

55
setup(
66
name='DeepInteract',
7-
version='1.0.9',
7+
version='1.1.0',
88
description='A geometric deep learning pipeline for predicting protein interface contacts.',
99
author='Alex Morehead',
1010
author_email='[email protected]',

0 commit comments

Comments
 (0)