Merge branch 'main' of github.com:inoue0426/scBiGCN

inoue0426 · Mar 31, 2024 · f572c15 · f572c15
2 parents 5aab9e9 + b21aa35
commit f572c15
Show file tree

Hide file tree

Showing 7 changed files with 105 additions and 56 deletions.
diff --git a/.github/workflows/python-format.yml b/.github/workflows/python-format.yml
@@ -0,0 +1,27 @@
+name: Python Formatting
+
+on: [push]
+
+jobs:
+  format-code:
+    runs-on: ubuntu-latest
+
+    permissions:
+      # Give the default GITHUB_TOKEN write permission to commit and push the
+      # added or changed files to the repository.
+      contents: write
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.10'
+    - name: Install black
+      run: python -m pip install black
+    - name: Format code
+      run: black .
+    - name: Commit changes
+      uses: stefanzweifel/git-auto-commit-action@v5
+      with:
+        commit_message: Apply auto-formatting
+        branch: ${{ github.head_ref }}
diff --git a/Dockerfile b/Dockerfile
@@ -0,0 +1,8 @@
+FROM python:3.9-slim
+USER root
+
+# install the notebook package
+RUN pip install --no-cache --upgrade pip && \
+    pip install --no-cache  torch --index-url https://download.pytorch.org/whl/cpu && \
+    pip install --no-cache  torch_geometric && \
+    pip install --no-cache  pyg_lib torch_sparse -f https://data.pyg.org/whl/torch-2.0.0+cpu.html
diff --git a/README.md b/README.md
@@ -1,40 +1,48 @@
 # BiGCN: Leveraging Cell and Gene Similarities for Single-cell Transcriptome Imputation with Bi-Graph Convolutional Networks
 
-This is an official implementation of the paper, "BiGCN: Leveraging Cell and Gene Similarities for Single-cell Transcriptome Imputation with Bi-Graph Convolutional Networks"
+![](overview.png)
 
-scBiGCN is a method that utilizes two GCNs to reconstruct gene expression matrices based on the similarity matrices of cells and genes. Utilizing each similarity enables the recovery of information lost through Dropout.
+This is an official implementation of the paper, "BiGCN: Leveraging Cell and Gene Similarities for Single-cell Transcriptome Imputation with Bi-Graph Convolutional Networks"
 
-<img width="1027" alt="Screenshot 2023-09-14 at 15 49 53" src="https://github.com/inoue0426/scBiGCN/assets/8393063/c9d1fbc0-bdf0-49b3-91b3-50181cbe16ec">
+scBiGCN is a method that utilizes two GCNs to reconstruct gene expression matrices based on the similarity matrices of cells and genes. Using each similarity enables the recovery of information lost through Dropout.
 
 scBiGCN has been implemented in Python.
 
 To get started immediately, check out our tutorials:
 - [Tutorial](https://github.com/inoue0426/scBiGCN/blob/main/sample%20notebook.ipynb)
 
-## Installation from GitHub
-To clone the repository and install manually, run the following from a terminal:
-```
-git clone [email protected]:inoue0426/scBiGCN.git
-cd scBiGCN
-conda create --name scBiGCN python=3.10 -y
-conda activate scBiGCN
-pip install -r requirement.txt
-```
-
 ## Requirement
 
 ```
 numpy==1.23.5
 pandas==2.0.3
 scikit-learn==1.3.0
-torch==1.13.1
+torch==1.13.1+cu116
 torch-geometric==2.3.1
-torch-sparse==0.6.17
+torch-sparse==0.6.17+pt113cu116
 tqdm==4.65.0
 ```
 
 ** Note: To utilize GPU acceleration, it is necessary to configure your own CUDA environment for PyTorch and PyTorch-Sparse.
 
+## Environment
+
+Our experiment was conducted on Ubuntu with an RTX 2080 and Nvidia A100. To use this model, please adjust the requirements to suit your environment, paying special attention to the PyTorch-related libraries. 
+
+## Installation from GitHub
+To clone the repository and install manually, run the following from a terminal:
+```
+git clone [email protected]:inoue0426/scBiGCN.git
+cd scBiGCN
+conda create --name scBiGCN python=3.10 -y
+conda activate scBiGCN
+pip install -r  requirement.txt
+# Please make sure to change the version to match the version of your GPU/CPU machine exactly.
+pip install --no-cache-dir  torch==1.13.1+cu116 --extra-index-url https://download.pytorch.org/whl/cu116
+pip install --no-cache-dir torch_geometric
+pip install --no-cache-dir pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-1.13.1%2Bcu116.html
+```
+
 ## Usage
 
 ### Quick Start
@@ -50,4 +58,4 @@ bigcn.run_model(df, verbose=True)
 ```
 
 ## Help
-If you have any questions or require assistance using MAGIC, please feel free to make a issues on https://github.com/inoue0426/scBiGCN/
+If you have any questions or require assistance using MAGIC, please feel free to make an issue on https://github.com/inoue0426/scBiGCN/
diff --git a/Tutorial.ipynb b/Tutorial.ipynb
@@ -2,10 +2,19 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 2,
    "id": "53d47961",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/export/scratch/inoue019/envs/scBiGCN/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n"
+     ]
+    }
+   ],
    "source": [
     "import pandas as pd\n",
     "import numpy as np\n",
@@ -32,7 +41,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 3,
    "id": "ede9e8c6",
    "metadata": {
     "scrolled": true
@@ -395,7 +404,7 @@
        "[1000 rows x 5000 columns]"
       ]
      },
-     "execution_count": 2,
+     "execution_count": 3,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -415,7 +424,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 4,
    "id": "bc4d19df",
    "metadata": {},
    "outputs": [
@@ -789,7 +798,7 @@
        "[1000 rows x 5000 columns]"
       ]
      },
-     "execution_count": 3,
+     "execution_count": 4,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -809,23 +818,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 5,
    "id": "486ba660",
    "metadata": {},
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "100%|██████████| 1500/1500 [02:01<00:00, 12.37it/s]"
+      "100%|██████████| 1500/1500 [02:01<00:00, 12.34it/s]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "CPU times: user 2min 6s, sys: 3.59 s, total: 2min 10s\n",
-      "Wall time: 2min 3s\n"
+      "CPU times: user 2min 6s, sys: 3.54 s, total: 2min 10s\n",
+      "Wall time: 2min 4s\n"
      ]
     },
     {
@@ -838,22 +847,22 @@
     {
      "data": {
       "text/plain": [
-       "array([[0.02243742, 0.        , 0.176742  , ..., 0.11798435, 0.        ,\n",
-       "        0.0194989 ],\n",
-       "       [0.16287151, 0.03038947, 0.12390558, ..., 0.04714939, 0.023552  ,\n",
-       "        0.25994456],\n",
-       "       [0.04037532, 0.01580523, 0.13696328, ..., 0.13242072, 0.        ,\n",
-       "        0.05790219],\n",
+       "array([[0.06184705, 0.01784644, 0.1660583 , ..., 0.02843533, 0.        ,\n",
+       "        0.03826597],\n",
+       "       [0.15132436, 0.03581575, 0.1416667 , ..., 0.0666969 , 0.00558404,\n",
+       "        0.22870545],\n",
+       "       [0.        , 0.        , 0.17596829, ..., 0.00705822, 0.        ,\n",
+       "        0.00371063],\n",
        "       ...,\n",
-       "       [0.07433029, 0.04123527, 0.10616948, ..., 0.04192581, 0.02231284,\n",
-       "        0.06234603],\n",
-       "       [0.08945477, 0.05060229, 0.1150779 , ..., 0.05662287, 0.0437512 ,\n",
-       "        0.07167498],\n",
-       "       [0.06099119, 0.        , 0.02700689, ..., 0.06171517, 0.0186478 ,\n",
-       "        0.11376322]], dtype=float32)"
+       "       [0.07470302, 0.02687185, 0.14962952, ..., 0.05768366, 0.01138911,\n",
+       "        0.03132814],\n",
+       "       [0.08545761, 0.03409339, 0.1452198 , ..., 0.06482822, 0.01634106,\n",
+       "        0.07690162],\n",
+       "       [0.08783356, 0.00831476, 0.03660469, ..., 0.03977096, 0.0350069 ,\n",
+       "        0.10716291]], dtype=float32)"
       ]
      },
-     "execution_count": 4,
+     "execution_count": 5,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -890,7 +899,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 6,
    "id": "bca9f626",
    "metadata": {
     "scrolled": true
@@ -1253,7 +1262,7 @@
        "[1000 rows x 5000 columns]"
       ]
      },
-     "execution_count": 5,
+     "execution_count": 6,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1273,7 +1282,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 7,
    "id": "92139141",
    "metadata": {},
    "outputs": [
@@ -1647,7 +1656,7 @@
        "[995 rows x 4744 columns]"
       ]
      },
-     "execution_count": 6,
+     "execution_count": 7,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1672,23 +1681,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
    "id": "4fb7ab66",
    "metadata": {},
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "100%|██████████| 100/100 [00:03<00:00, 25.88it/s]"
+      "100%|██████████| 100/100 [00:03<00:00, 26.11it/s]"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "CPU times: user 8.35 s, sys: 2.53 s, total: 10.9 s\n",
-      "Wall time: 4.9 s\n"
+      "CPU times: user 8.22 s, sys: 2.67 s, total: 10.9 s\n",
+      "Wall time: 4.87 s\n"
      ]
     },
     {
@@ -1716,7 +1725,7 @@
        "        -1.1651450e-02, -1.5314183e-02, -2.4916604e-03]], dtype=float32)"
       ]
      },
-     "execution_count": 7,
+     "execution_count": 8,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1738,9 +1747,9 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "test",
+   "display_name": "scBiGCN",
    "language": "python",
-   "name": "test"
+   "name": "scbigcn"
   },
   "language_info": {
    "codemirror_mode": {

diff --git a/bigcn.py b/bigcn.py
@@ -84,7 +84,7 @@ def forward(self, data, x, adj, x_t, adj_t, clustering):
         return res
 
 
-def run_model(input_data, params=None, clustering=False, verbose=False):
+def run_model(input_data, params=None, clustering=False, verbose=False, device=False):
     """Run model
 
     input_data: gene expression matrix
@@ -114,8 +114,8 @@ def run_model(input_data, params=None, clustering=False, verbose=False):
             "optimizer": "Adam",
             "clustering": True,
         }
-
-    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+    if device:
+        device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
     x, adj = get_data(input_data)
     x_t, adj_t = get_data(input_data.T)

diff --git a/overview.png b/overview.png
diff --git a/requirement.txt b/requirement.txt
@@ -1,7 +1,4 @@
 numpy==1.23.5
 pandas==2.0.3
 scikit-learn==1.3.0
-torch==1.13.1
-torch-geometric==2.3.1
-torch-sparse==0.6.17
 tqdm==4.65.0