Upload app.py

app.py

@@ -0,0 +1,107 @@
+import gradio as gr
+import os,uuid,pysam
+import time,gdown
+import pyBigWig,pickle
+from scipy.sparse import csr_matrix
+import numpy as np
+
+def atac_bwtonpz(atac_file):
+    bw = pyBigWig.open(atac_file)
+    signals = {}
+    for chrom, length in bw.chroms().items():
+        try:
+            if chrom == 'chrX':
+                chr = 'X'
+            else:
+                chr = int(chrom[3:])
+        except Exception:
+            continue
+        temp = np.zeros(length)
+        intervals = bw.intervals(chrom)
+        for interval in intervals:
+            temp[interval[0]:interval[1]] = interval[2]
+
+        seq_length = length // 1000 * 1000
+        signals[chr] = csr_matrix(temp.astype('float32')[:seq_length])
+    with open('ATAC/atac_'+atac_file.replace('bigWig','pickle'),'wb') as f:
+        pickle.dump(signals,f)
+
+
+def process(bam_file,progress=gr.Progress()):
+    if not os.path.exists('ATAC'):
+        os.mkdir('ATAC')
+    else:
+        for f in os.listdir('ATAC/'):
+            if not f.startswith('example'):
+                os.remove(os.path.join('ATAC/', f))
+    if bam_file=='':
+        raise gr.Error('')
+
+    fid = str(uuid.uuid4())
+    print(bam_file,fid)
+
+    if 'dropbox' in bam_file:
+        progress(0.2, desc="Downloading BAM file "+bam_file)
+        os.system('wget -O ATAC/'+fid+'.bam '+bam_file)
+    else:
+        progress(0.2, desc="Downloading BAM file "+bam_file)
+        gdown.download(url=bam_file,fuzzy=True, output='ATAC/'+fid + '.bam')
+        # tmp_bamid=bam_file.split('file/d/')[1].split('/view?')[0]
+        # gdown.download(id=tmp_bamid,output=fid + '.bam')
+    progress(0.4, desc="Index BAM file")
+    time.sleep(0.1)
+    pysam.index('ATAC/'+fid + '.bam')
+    progress(0.6, desc="Convert BAM file to bigWig file")
+    time.sleep(0.1)
+    os.system('bamCoverage --bam %s -o %s --outFileFormat bigwig --normalizeUsing RPGC '
+              '--effectiveGenomeSize 2913022398 --Offset 1 --binSize 1 --numberOfProcessors 6 '
+              '--blackListFileName data/black_list.bed' % ('ATAC/'+fid + '.bam','ATAC/'+ fid + '.bigWig'))
+    progress(0.8, desc="Process bigWig file")
+    time.sleep(0.1)
+    atac_bwtonpz('ATAC/'+fid + '.bigWig')
+    os.remove('ATAC/'+fid+'.bam.bai')
+    os.remove('ATAC/'+fid + '.bigWig')
+    os.remove('ATAC/'+fid + '.bam')
+    return 'ATAC/atac_' + fid + '.pickle'
+
+
+css='li {font-size: 15px;}'
+
+with gr.Blocks(theme=gr.themes.Monochrome(),css=css) as app:
+    gr.Markdown('<h1 style="text-align: center; font-size: 2.2em"> '
+                'A computational tool to use ATAC-seq to impute epigenome, transcriptome, and high-resolution chromatin contact maps</h1>')
+
+    gr.HTML('<div style="text-align: center;">'
+            '<a href="https://github.com/zzh24zzh/EPCOT_gradio" style="margin-right: 20px;font-size: 18px; color:black;">[Code]</a>'
+            '<a href="https://www.biorxiv.org/content/10.1101/2022.05.23.493129v2" style="font-size: 18px; color:black;">[Paper]</a>'
+            '</div>')
+    # gr.Markdown('[[Code]](https://github.com/zzh24zzh/EPCOT_gradio) [[Paper]](https://www.biorxiv.org/content/10.1101/2022.05.23.493129v2)')
+    gr.Markdown('## Instructions:\n'
+                '- Use the first app to process ATAC-seq data by uploading an ATAC-seq BAM file.\n'
+                '- Execute the model by uploading the processed ATAC-seq file to the "Run Model" interface in the second app.\n'
+                '- Visualize the prediction results by uploading the predicted files from the "Run Model" interface to the "Visualize Prediction Results" interface.\n')
+    gr.Markdown('<hr>')
+    gr.Markdown('<h1 font-size: 2em"> App 1: Process ATAC-seq data</h1>\n'
+                # '<p style="font-size: 15px;">Hosted on a server from Liu Lab</p>\n'
+                '<p style="font-size: 15px;"><strong>Notice:</strong> Downsample the BAM file before upload if the size is too large</p>')
+    with gr.Row():
+        with gr.Column():
+            inp=gr.Textbox(
+                label="URL to ATAC-seq BAM",
+                info='Only Dropbox and Google Drive file links are accecpted (set "Anyone with the link")',
+                lines=1,
+            )
+            btn2=gr.Button("Submit")
+        with gr.Column():
+            out = gr.File(label='Download the processed file')
+        btn2.click(fn=process, inputs=inp, outputs=out)
+
+    gr.Markdown('<hr>')
+    gr.Markdown('<h1 font-size: 2em;"> App 2: Run models and visualize prediction results</h1>\n'
+                '<p style="font-size: 16px;">'
+            'Hosted on <a href="https://huggingface.co/spaces/drjieliu/EPCOT" style="font-size: 16px;">Hugging Face Space</a> </p>')
+    with gr.Column():
+        gr.HTML('<iframe src="https://drjieliu-epcot.hf.space" '
+                'border="none" width="100%" height=1200></iframe>')
+
+app.launch(debug=True,enable_queue=True)