Spaces:
Sleeping
Sleeping
supercat666
commited on
Commit
·
fc8ed8c
1
Parent(s):
fd276e2
fix app
Browse files- app.py +246 -151
- cas12lstmvcf.py +68 -8
app.py
CHANGED
@@ -5,6 +5,7 @@ import cas9attvcf
|
|
5 |
import cas9off
|
6 |
import cas12
|
7 |
import cas12lstm
|
|
|
8 |
import pandas as pd
|
9 |
import streamlit as st
|
10 |
import plotly.graph_objs as go
|
@@ -26,7 +27,7 @@ CRISPR_MODELS = ['Cas9', 'Cas12', 'Cas13d']
|
|
26 |
|
27 |
selected_model = st.selectbox('Select CRISPR model:', CRISPR_MODELS, key='selected_model')
|
28 |
cas9att_path = 'cas9_model/Cas9_MultiHeadAttention_weights.h5'
|
29 |
-
|
30 |
|
31 |
#plot functions
|
32 |
def generate_coolbox_plot(bigwig_path, region, output_image_path):
|
@@ -331,7 +332,7 @@ if selected_model == 'Cas9':
|
|
331 |
# Process predictions
|
332 |
if predict_button and gene_symbol:
|
333 |
with st.spinner('Predicting... Please wait'):
|
334 |
-
predictions, gene_sequence, exons = cas9attvcf.process_gene(gene_symbol, cas9att_path)
|
335 |
full_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)
|
336 |
sorted_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)[:10]
|
337 |
st.session_state['full_results'] = full_predictions
|
@@ -489,6 +490,11 @@ if selected_model == 'Cas9':
|
|
489 |
st.experimental_rerun()
|
490 |
|
491 |
elif selected_model == 'Cas12':
|
|
|
|
|
|
|
|
|
|
|
492 |
# Gene symbol entry with autocomplete-like feature
|
493 |
gene_symbol = st.selectbox('Enter a Gene Symbol:', [''] + gene_symbol_list, key='gene_symbol',
|
494 |
format_func=lambda x: x if x else "")
|
@@ -497,159 +503,248 @@ elif selected_model == 'Cas12':
|
|
497 |
if 'current_gene_symbol' not in st.session_state:
|
498 |
st.session_state['current_gene_symbol'] = ""
|
499 |
|
500 |
-
|
501 |
-
|
502 |
-
|
503 |
-
# Function to clean up old files
|
504 |
-
def clean_up_old_files(gene_symbol):
|
505 |
-
genbank_file_path = f"{gene_symbol}_crispr_targets.gb"
|
506 |
-
bed_file_path = f"{gene_symbol}_crispr_targets.bed"
|
507 |
-
csv_file_path = f"{gene_symbol}_crispr_predictions.csv"
|
508 |
-
for path in [genbank_file_path, bed_file_path, csv_file_path]:
|
509 |
-
if os.path.exists(path):
|
510 |
-
os.remove(path)
|
511 |
|
512 |
-
|
513 |
-
|
514 |
-
clean_up_old_files(st.session_state['current_gene_symbol'])
|
515 |
-
|
516 |
-
# Process predictions
|
517 |
-
if predict_button and gene_symbol:
|
518 |
-
with st.spinner('Predicting... Please wait'):
|
519 |
-
predictions, gene_sequence, exons = cas12lstm.process_gene(gene_symbol, cas9att_path)
|
520 |
-
sorted_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)[:10]
|
521 |
-
st.session_state['on_target_results'] = sorted_predictions
|
522 |
-
st.session_state['gene_sequence'] = gene_sequence # Save gene sequence in session state
|
523 |
-
st.session_state['exons'] = exons # Store exon data
|
524 |
-
|
525 |
-
# Notify the user once the process is completed successfully.
|
526 |
-
st.success('Prediction completed!')
|
527 |
-
st.session_state['prediction_made'] = True
|
528 |
-
|
529 |
-
if 'on_target_results' in st.session_state and st.session_state['on_target_results']:
|
530 |
-
ensembl_id = gene_annotations.get(gene_symbol, 'Unknown') # Get Ensembl ID or default to 'Unknown'
|
531 |
-
col1, col2, col3 = st.columns(3)
|
532 |
-
with col1:
|
533 |
-
st.markdown("**Genome**")
|
534 |
-
st.markdown("Homo sapiens")
|
535 |
-
with col2:
|
536 |
-
st.markdown("**Gene**")
|
537 |
-
st.markdown(f"{gene_symbol} : {ensembl_id} (primary)")
|
538 |
-
with col3:
|
539 |
-
st.markdown("**Nuclease**")
|
540 |
-
st.markdown("SpCas9")
|
541 |
-
# Include "Target" in the DataFrame's columns
|
542 |
-
try:
|
543 |
-
df = pd.DataFrame(st.session_state['on_target_results'],
|
544 |
-
columns=["Chr", "Start Pos", "End Pos", "Strand", "Transcript", "Exon", "Target",
|
545 |
-
"gRNA", "Prediction"])
|
546 |
-
st.dataframe(df)
|
547 |
-
except ValueError as e:
|
548 |
-
st.error(f"DataFrame creation error: {e}")
|
549 |
-
# Optionally print or log the problematic data for debugging:
|
550 |
-
print(st.session_state['on_target_results'])
|
551 |
-
|
552 |
-
# Initialize Plotly figure
|
553 |
-
fig = go.Figure()
|
554 |
-
|
555 |
-
EXON_BASE = 0 # Base position for exons and CDS on the Y axis
|
556 |
-
EXON_HEIGHT = 0.02 # How 'tall' the exon markers should appear
|
557 |
-
|
558 |
-
# Plot Exons as small markers on the X-axis
|
559 |
-
for exon in st.session_state['exons']:
|
560 |
-
exon_start, exon_end = exon['start'], exon['end']
|
561 |
-
fig.add_trace(go.Bar(
|
562 |
-
x=[(exon_start + exon_end) / 2],
|
563 |
-
y=[EXON_HEIGHT],
|
564 |
-
width=[exon_end - exon_start],
|
565 |
-
base=EXON_BASE,
|
566 |
-
marker_color='rgba(128, 0, 128, 0.5)',
|
567 |
-
name='Exon'
|
568 |
-
))
|
569 |
-
|
570 |
-
VERTICAL_GAP = 0.2 # Gap between different ranks
|
571 |
-
|
572 |
-
# Define max and min Y values based on strand and rank
|
573 |
-
MAX_STRAND_Y = 0.1 # Maximum Y value for positive strand results
|
574 |
-
MIN_STRAND_Y = -0.1 # Minimum Y value for negative strand results
|
575 |
-
|
576 |
-
# Iterate over top 5 sorted predictions to create the plot
|
577 |
-
for i, prediction in enumerate(st.session_state['on_target_results'][:5], start=1): # Only top 5
|
578 |
-
chrom, start, end, strand, transcript, exon, target, gRNA, prediction_score = prediction
|
579 |
-
midpoint = (int(start) + int(end)) / 2
|
580 |
-
|
581 |
-
# Vertical position based on rank, modified by strand
|
582 |
-
y_value = (MAX_STRAND_Y - (i - 1) * VERTICAL_GAP) if strand == '1' or strand == '+' else (
|
583 |
-
MIN_STRAND_Y + (i - 1) * VERTICAL_GAP)
|
584 |
-
|
585 |
-
fig.add_trace(go.Scatter(
|
586 |
-
x=[midpoint],
|
587 |
-
y=[y_value],
|
588 |
-
mode='markers+text',
|
589 |
-
marker=dict(symbol='triangle-up' if strand == '1' or strand == '+' else 'triangle-down',
|
590 |
-
size=12),
|
591 |
-
text=f"Rank: {i}", # Text label
|
592 |
-
hoverinfo='text',
|
593 |
-
hovertext=f"Rank: {i}<br>Chromosome: {chrom}<br>Target Sequence: {target}<br>gRNA: {gRNA}<br>Start: {start}<br>End: {end}<br>Strand: {'+' if strand == '1' or strand == '+' else '-'}<br>Transcript: {transcript}<br>Prediction: {prediction_score:.4f}",
|
594 |
-
))
|
595 |
-
|
596 |
-
# Update layout for clarity and interaction
|
597 |
-
fig.update_layout(
|
598 |
-
title='Top 5 gRNA Sequences by Prediction Score',
|
599 |
-
xaxis_title='Genomic Position',
|
600 |
-
yaxis_title='Strand',
|
601 |
-
yaxis=dict(tickvals=[MAX_STRAND_Y, MIN_STRAND_Y], ticktext=['+', '-']),
|
602 |
-
showlegend=False,
|
603 |
-
hovermode='x unified',
|
604 |
-
)
|
605 |
-
|
606 |
-
# Display the plot
|
607 |
-
st.plotly_chart(fig)
|
608 |
-
|
609 |
-
if 'gene_sequence' in st.session_state and st.session_state['gene_sequence']:
|
610 |
-
gene_symbol = st.session_state['current_gene_symbol']
|
611 |
-
gene_sequence = st.session_state['gene_sequence']
|
612 |
-
|
613 |
-
# Define file paths
|
614 |
genbank_file_path = f"{gene_symbol}_crispr_targets.gb"
|
615 |
bed_file_path = f"{gene_symbol}_crispr_targets.bed"
|
616 |
csv_file_path = f"{gene_symbol}_crispr_predictions.csv"
|
617 |
-
|
618 |
-
|
619 |
-
|
620 |
-
|
621 |
-
|
622 |
-
|
623 |
-
|
624 |
-
|
625 |
-
|
626 |
-
|
627 |
-
|
628 |
-
|
629 |
-
|
630 |
-
|
631 |
-
|
632 |
-
|
633 |
-
|
634 |
-
|
635 |
-
#
|
636 |
-
|
637 |
-
|
638 |
-
|
639 |
-
|
640 |
-
|
641 |
-
|
642 |
-
|
643 |
-
|
644 |
-
|
645 |
-
|
646 |
-
|
647 |
-
|
648 |
-
|
649 |
-
|
650 |
-
|
651 |
-
|
652 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
653 |
|
654 |
elif selected_model == 'Cas13d':
|
655 |
ENTRY_METHODS = dict(
|
|
|
5 |
import cas9off
|
6 |
import cas12
|
7 |
import cas12lstm
|
8 |
+
import cas12lstmvcf
|
9 |
import pandas as pd
|
10 |
import streamlit as st
|
11 |
import plotly.graph_objs as go
|
|
|
27 |
|
28 |
selected_model = st.selectbox('Select CRISPR model:', CRISPR_MODELS, key='selected_model')
|
29 |
cas9att_path = 'cas9_model/Cas9_MultiHeadAttention_weights.h5'
|
30 |
+
cas12lstm_path = 'cas12_model/BiLSTM_Cpf1_weights.h5'
|
31 |
|
32 |
#plot functions
|
33 |
def generate_coolbox_plot(bigwig_path, region, output_image_path):
|
|
|
332 |
# Process predictions
|
333 |
if predict_button and gene_symbol:
|
334 |
with st.spinner('Predicting... Please wait'):
|
335 |
+
predictions, gene_sequence, exons = cas9attvcf.process_gene(gene_symbol, vcf_reader, cas9att_path)
|
336 |
full_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)
|
337 |
sorted_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)[:10]
|
338 |
st.session_state['full_results'] = full_predictions
|
|
|
490 |
st.experimental_rerun()
|
491 |
|
492 |
elif selected_model == 'Cas12':
|
493 |
+
cas12target_selection = st.radio(
|
494 |
+
"Select either mutation or not:",
|
495 |
+
('regular', 'mutation'),
|
496 |
+
key='cas12target_selection'
|
497 |
+
)
|
498 |
# Gene symbol entry with autocomplete-like feature
|
499 |
gene_symbol = st.selectbox('Enter a Gene Symbol:', [''] + gene_symbol_list, key='gene_symbol',
|
500 |
format_func=lambda x: x if x else "")
|
|
|
503 |
if 'current_gene_symbol' not in st.session_state:
|
504 |
st.session_state['current_gene_symbol'] = ""
|
505 |
|
506 |
+
if cas12target_selection == 'regular':
|
507 |
+
# Prediction button
|
508 |
+
predict_button = st.button('Predict on-target')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
509 |
|
510 |
+
# Function to clean up old files
|
511 |
+
def clean_up_old_files(gene_symbol):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
512 |
genbank_file_path = f"{gene_symbol}_crispr_targets.gb"
|
513 |
bed_file_path = f"{gene_symbol}_crispr_targets.bed"
|
514 |
csv_file_path = f"{gene_symbol}_crispr_predictions.csv"
|
515 |
+
for path in [genbank_file_path, bed_file_path, csv_file_path]:
|
516 |
+
if os.path.exists(path):
|
517 |
+
os.remove(path)
|
518 |
+
|
519 |
+
|
520 |
+
# Clean up files if a new gene symbol is entered
|
521 |
+
if st.session_state['current_gene_symbol'] and gene_symbol != st.session_state['current_gene_symbol']:
|
522 |
+
clean_up_old_files(st.session_state['current_gene_symbol'])
|
523 |
+
|
524 |
+
# Process predictions
|
525 |
+
if predict_button and gene_symbol:
|
526 |
+
with st.spinner('Predicting... Please wait'):
|
527 |
+
predictions, gene_sequence, exons = cas12lstm.process_gene(gene_symbol, cas12lstm_path)
|
528 |
+
sorted_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)[:10]
|
529 |
+
st.session_state['on_target_results'] = sorted_predictions
|
530 |
+
st.session_state['gene_sequence'] = gene_sequence # Save gene sequence in session state
|
531 |
+
st.session_state['exons'] = exons # Store exon data
|
532 |
+
|
533 |
+
# Notify the user once the process is completed successfully.
|
534 |
+
st.success('Prediction completed!')
|
535 |
+
st.session_state['prediction_made'] = True
|
536 |
+
|
537 |
+
if 'on_target_results' in st.session_state and st.session_state['on_target_results']:
|
538 |
+
ensembl_id = gene_annotations.get(gene_symbol, 'Unknown') # Get Ensembl ID or default to 'Unknown'
|
539 |
+
col1, col2, col3 = st.columns(3)
|
540 |
+
with col1:
|
541 |
+
st.markdown("**Genome**")
|
542 |
+
st.markdown("Homo sapiens")
|
543 |
+
with col2:
|
544 |
+
st.markdown("**Gene**")
|
545 |
+
st.markdown(f"{gene_symbol} : {ensembl_id} (primary)")
|
546 |
+
with col3:
|
547 |
+
st.markdown("**Nuclease**")
|
548 |
+
st.markdown("SpCas9")
|
549 |
+
# Include "Target" in the DataFrame's columns
|
550 |
+
try:
|
551 |
+
df = pd.DataFrame(st.session_state['on_target_results'],
|
552 |
+
columns=["Chr", "Start Pos", "End Pos", "Strand", "Transcript", "Exon",
|
553 |
+
"Target",
|
554 |
+
"gRNA", "Prediction"])
|
555 |
+
st.dataframe(df)
|
556 |
+
except ValueError as e:
|
557 |
+
st.error(f"DataFrame creation error: {e}")
|
558 |
+
# Optionally print or log the problematic data for debugging:
|
559 |
+
print(st.session_state['on_target_results'])
|
560 |
+
|
561 |
+
# Initialize Plotly figure
|
562 |
+
fig = go.Figure()
|
563 |
+
|
564 |
+
EXON_BASE = 0 # Base position for exons and CDS on the Y axis
|
565 |
+
EXON_HEIGHT = 0.02 # How 'tall' the exon markers should appear
|
566 |
+
|
567 |
+
# Plot Exons as small markers on the X-axis
|
568 |
+
for exon in st.session_state['exons']:
|
569 |
+
exon_start, exon_end = exon['start'], exon['end']
|
570 |
+
fig.add_trace(go.Bar(
|
571 |
+
x=[(exon_start + exon_end) / 2],
|
572 |
+
y=[EXON_HEIGHT],
|
573 |
+
width=[exon_end - exon_start],
|
574 |
+
base=EXON_BASE,
|
575 |
+
marker_color='rgba(128, 0, 128, 0.5)',
|
576 |
+
name='Exon'
|
577 |
+
))
|
578 |
+
|
579 |
+
VERTICAL_GAP = 0.2 # Gap between different ranks
|
580 |
+
|
581 |
+
# Define max and min Y values based on strand and rank
|
582 |
+
MAX_STRAND_Y = 0.1 # Maximum Y value for positive strand results
|
583 |
+
MIN_STRAND_Y = -0.1 # Minimum Y value for negative strand results
|
584 |
+
|
585 |
+
# Iterate over top 5 sorted predictions to create the plot
|
586 |
+
for i, prediction in enumerate(st.session_state['on_target_results'][:5], start=1): # Only top 5
|
587 |
+
chrom, start, end, strand, transcript, exon, target, gRNA, prediction_score = prediction
|
588 |
+
midpoint = (int(start) + int(end)) / 2
|
589 |
+
|
590 |
+
# Vertical position based on rank, modified by strand
|
591 |
+
y_value = (MAX_STRAND_Y - (i - 1) * VERTICAL_GAP) if strand == '1' or strand == '+' else (
|
592 |
+
MIN_STRAND_Y + (i - 1) * VERTICAL_GAP)
|
593 |
+
|
594 |
+
fig.add_trace(go.Scatter(
|
595 |
+
x=[midpoint],
|
596 |
+
y=[y_value],
|
597 |
+
mode='markers+text',
|
598 |
+
marker=dict(symbol='triangle-up' if strand == '1' or strand == '+' else 'triangle-down',
|
599 |
+
size=12),
|
600 |
+
text=f"Rank: {i}", # Text label
|
601 |
+
hoverinfo='text',
|
602 |
+
hovertext=f"Rank: {i}<br>Chromosome: {chrom}<br>Target Sequence: {target}<br>gRNA: {gRNA}<br>Start: {start}<br>End: {end}<br>Strand: {'+' if strand == '1' or strand == '+' else '-'}<br>Transcript: {transcript}<br>Prediction: {prediction_score:.4f}",
|
603 |
+
))
|
604 |
+
|
605 |
+
# Update layout for clarity and interaction
|
606 |
+
fig.update_layout(
|
607 |
+
title='Top 5 gRNA Sequences by Prediction Score',
|
608 |
+
xaxis_title='Genomic Position',
|
609 |
+
yaxis_title='Strand',
|
610 |
+
yaxis=dict(tickvals=[MAX_STRAND_Y, MIN_STRAND_Y], ticktext=['+', '-']),
|
611 |
+
showlegend=False,
|
612 |
+
hovermode='x unified',
|
613 |
+
)
|
614 |
+
|
615 |
+
# Display the plot
|
616 |
+
st.plotly_chart(fig)
|
617 |
+
|
618 |
+
if 'gene_sequence' in st.session_state and st.session_state['gene_sequence']:
|
619 |
+
gene_symbol = st.session_state['current_gene_symbol']
|
620 |
+
gene_sequence = st.session_state['gene_sequence']
|
621 |
+
|
622 |
+
# Define file paths
|
623 |
+
genbank_file_path = f"{gene_symbol}_crispr_targets.gb"
|
624 |
+
bed_file_path = f"{gene_symbol}_crispr_targets.bed"
|
625 |
+
csv_file_path = f"{gene_symbol}_crispr_predictions.csv"
|
626 |
+
plot_image_path = f"{gene_symbol}_gtracks_plot.png"
|
627 |
+
|
628 |
+
# Generate files
|
629 |
+
cas12lstm.generate_genbank_file_from_df(df, gene_sequence, gene_symbol, genbank_file_path)
|
630 |
+
cas12lstm.create_bed_file_from_df(df, bed_file_path)
|
631 |
+
cas12lstm.create_csv_from_df(df, csv_file_path)
|
632 |
+
|
633 |
+
# Prepare an in-memory buffer for the ZIP file
|
634 |
+
zip_buffer = io.BytesIO()
|
635 |
+
with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
|
636 |
+
# For each file, add it to the ZIP file
|
637 |
+
zip_file.write(genbank_file_path)
|
638 |
+
zip_file.write(bed_file_path)
|
639 |
+
zip_file.write(csv_file_path)
|
640 |
+
|
641 |
+
# Important: move the cursor to the beginning of the BytesIO buffer before reading it
|
642 |
+
zip_buffer.seek(0)
|
643 |
+
|
644 |
+
# Specify the region you want to visualize
|
645 |
+
min_start = df['Start Pos'].min()
|
646 |
+
max_end = df['End Pos'].max()
|
647 |
+
chromosome = df['Chr'].mode()[0] # Assumes most common chromosome is the target
|
648 |
+
region = f"{chromosome}:{min_start}-{max_end}"
|
649 |
+
|
650 |
+
# Generate the pyGenomeTracks plot
|
651 |
+
gtracks_command = f"gtracks {region} {bed_file_path} {plot_image_path}"
|
652 |
+
subprocess.run(gtracks_command, shell=True)
|
653 |
+
st.image(plot_image_path)
|
654 |
+
|
655 |
+
# Display the download button for the ZIP file
|
656 |
+
st.download_button(
|
657 |
+
label="Download GenBank, BED, CSV files as ZIP",
|
658 |
+
data=zip_buffer.getvalue(),
|
659 |
+
file_name=f"{gene_symbol}_files.zip",
|
660 |
+
mime="application/zip"
|
661 |
+
)
|
662 |
+
elif cas12target_selection == 'mutation':
|
663 |
+
# Prediction button
|
664 |
+
predict_button = st.button('Predict on-target')
|
665 |
+
vcf_reader = cyvcf2.VCF('SRR25934512.filter.snps.indels.vcf.gz')
|
666 |
+
|
667 |
+
if 'exons' not in st.session_state:
|
668 |
+
st.session_state['exons'] = []
|
669 |
+
|
670 |
+
# Process predictions
|
671 |
+
if predict_button and gene_symbol:
|
672 |
+
with st.spinner('Predicting... Please wait'):
|
673 |
+
predictions, gene_sequence, exons = cas12lstmvcf.process_gene(gene_symbol, vcf_reader,
|
674 |
+
cas12lstm_path)
|
675 |
+
full_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)
|
676 |
+
sorted_predictions = sorted(predictions, key=lambda x: x[8], reverse=True)[:10]
|
677 |
+
st.session_state['full_results'] = full_predictions
|
678 |
+
st.session_state['on_target_results'] = sorted_predictions
|
679 |
+
st.session_state['gene_sequence'] = gene_sequence # Save gene sequence in session state
|
680 |
+
st.session_state['exons'] = exons # Store exon data
|
681 |
+
|
682 |
+
# Notify the user once the process is completed successfully.
|
683 |
+
st.success('Prediction completed!')
|
684 |
+
st.session_state['prediction_made'] = True
|
685 |
+
|
686 |
+
if 'on_target_results' in st.session_state and st.session_state['on_target_results']:
|
687 |
+
ensembl_id = gene_annotations.get(gene_symbol,
|
688 |
+
'Unknown') # Get Ensembl ID or default to 'Unknown'
|
689 |
+
col1, col2, col3 = st.columns(3)
|
690 |
+
with col1:
|
691 |
+
st.markdown("**Genome**")
|
692 |
+
st.markdown("Homo sapiens")
|
693 |
+
with col2:
|
694 |
+
st.markdown("**Gene**")
|
695 |
+
st.markdown(f"{gene_symbol} : {ensembl_id} (primary)")
|
696 |
+
with col3:
|
697 |
+
st.markdown("**Nuclease**")
|
698 |
+
st.markdown("SpCas9")
|
699 |
+
# Include "Target" in the DataFrame's columns
|
700 |
+
try:
|
701 |
+
df = pd.DataFrame(st.session_state['on_target_results'],
|
702 |
+
columns=["Gene Symbol", "Chr", "Strand", "Target Start", "Transcript",
|
703 |
+
"Exon",
|
704 |
+
"Target",
|
705 |
+
"gRNA", "Prediction", "Is Mutation"])
|
706 |
+
df_full = pd.DataFrame(st.session_state['full_results'],
|
707 |
+
columns=["Gene Symbol", "Chr", "Strand", "Target Start",
|
708 |
+
"Transcript",
|
709 |
+
"Exon", "Target",
|
710 |
+
"gRNA", "Prediction", "Is Mutation"])
|
711 |
+
st.dataframe(df)
|
712 |
+
except ValueError as e:
|
713 |
+
st.error(f"DataFrame creation error: {e}")
|
714 |
+
# Optionally print or log the problematic data for debugging:
|
715 |
+
print(st.session_state['on_target_results'])
|
716 |
+
|
717 |
+
if 'gene_sequence' in st.session_state and st.session_state['gene_sequence']:
|
718 |
+
gene_symbol = st.session_state['current_gene_symbol']
|
719 |
+
gene_sequence = st.session_state['gene_sequence']
|
720 |
+
|
721 |
+
# Define file paths
|
722 |
+
genbank_file_path = f"{gene_symbol}_crispr_targets.gb"
|
723 |
+
bed_file_path = f"{gene_symbol}_crispr_targets.bed"
|
724 |
+
csv_file_path = f"{gene_symbol}_crispr_predictions.csv"
|
725 |
+
plot_image_path = f"{gene_symbol}_gtracks_plot.png"
|
726 |
+
|
727 |
+
# Generate files
|
728 |
+
cas12lstmvcf.generate_genbank_file_from_df(df_full, gene_sequence, gene_symbol,
|
729 |
+
genbank_file_path)
|
730 |
+
cas12lstmvcf.create_bed_file_from_df(df_full, bed_file_path)
|
731 |
+
cas12lstmvcf.create_csv_from_df(df_full, csv_file_path)
|
732 |
+
|
733 |
+
# Prepare an in-memory buffer for the ZIP file
|
734 |
+
zip_buffer = io.BytesIO()
|
735 |
+
with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
|
736 |
+
# For each file, add it to the ZIP file
|
737 |
+
zip_file.write(genbank_file_path)
|
738 |
+
zip_file.write(bed_file_path)
|
739 |
+
zip_file.write(csv_file_path)
|
740 |
+
|
741 |
+
# Display the download button for the ZIP file
|
742 |
+
st.download_button(
|
743 |
+
label="Download GenBank, BED, CSV files as ZIP",
|
744 |
+
data=zip_buffer.getvalue(),
|
745 |
+
file_name=f"{gene_symbol}_files.zip",
|
746 |
+
mime="application/zip"
|
747 |
+
)
|
748 |
|
749 |
elif selected_model == 'Cas13d':
|
750 |
ENTRY_METHODS = dict(
|
cas12lstmvcf.py
CHANGED
@@ -8,6 +8,10 @@ from keras.metrics import MeanSquaredError
|
|
8 |
|
9 |
import pandas as pd
|
10 |
import numpy as np
|
|
|
|
|
|
|
|
|
11 |
|
12 |
import requests
|
13 |
from functools import reduce
|
@@ -278,14 +282,70 @@ def process_gene(gene_symbol, vcf_reader, model_path):
|
|
278 |
print(f"Failed to retrieve gene sequence for exon {exon_id}.")
|
279 |
else:
|
280 |
print("Failed to retrieve transcripts.")
|
281 |
-
|
282 |
-
output = []
|
283 |
-
for result in results:
|
284 |
-
for item in result:
|
285 |
-
output.append(item)
|
286 |
-
# Sort results based on prediction score (assuming score is at the 8th index)
|
287 |
-
sorted_results = sorted(output, key=lambda x: x[8], reverse=True)
|
288 |
|
289 |
# Return the sorted output, combined gene sequences, and all exons
|
290 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
291 |
|
|
|
8 |
|
9 |
import pandas as pd
|
10 |
import numpy as np
|
11 |
+
from Bio import SeqIO
|
12 |
+
from Bio.SeqRecord import SeqRecord
|
13 |
+
from Bio.SeqFeature import SeqFeature, FeatureLocation
|
14 |
+
from Bio.Seq import Seq
|
15 |
|
16 |
import requests
|
17 |
from functools import reduce
|
|
|
282 |
print(f"Failed to retrieve gene sequence for exon {exon_id}.")
|
283 |
else:
|
284 |
print("Failed to retrieve transcripts.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
285 |
|
286 |
# Return the sorted output, combined gene sequences, and all exons
|
287 |
+
return results, all_gene_sequences, all_exons
|
288 |
+
|
289 |
+
def create_genbank_features(data):
|
290 |
+
features = []
|
291 |
+
|
292 |
+
# If the input data is a DataFrame, convert it to a list of lists
|
293 |
+
if isinstance(data, pd.DataFrame):
|
294 |
+
formatted_data = data.values.tolist()
|
295 |
+
elif isinstance(data, list):
|
296 |
+
formatted_data = data
|
297 |
+
else:
|
298 |
+
raise TypeError("Data should be either a list or a pandas DataFrame.")
|
299 |
+
|
300 |
+
for row in formatted_data:
|
301 |
+
try:
|
302 |
+
start = int(row[1])
|
303 |
+
end = start + len(row[6]) # Calculate the end position based on the target sequence length
|
304 |
+
except ValueError as e:
|
305 |
+
print(f"Error converting start/end to int: {row[1]}, {row[2]} - {e}")
|
306 |
+
continue
|
307 |
+
|
308 |
+
strand = 1 if row[3] == '1' else -1
|
309 |
+
location = FeatureLocation(start=start, end=end, strand=strand)
|
310 |
+
is_mutation = 'Yes' if row[9] else 'No'
|
311 |
+
feature = SeqFeature(location=location, type="misc_feature", qualifiers={
|
312 |
+
'label': row[7], # Use gRNA as the label
|
313 |
+
'note': f"Prediction: {row[8]}, Mutation: {is_mutation}" # Include the prediction score and mutation status
|
314 |
+
})
|
315 |
+
features.append(feature)
|
316 |
+
|
317 |
+
return features
|
318 |
+
|
319 |
+
def generate_genbank_file_from_df(df, gene_sequence, gene_symbol, output_path):
|
320 |
+
# Ensure gene_sequence is a string before creating Seq object
|
321 |
+
if not isinstance(gene_sequence, str):
|
322 |
+
gene_sequence = str(gene_sequence)
|
323 |
+
|
324 |
+
features = create_genbank_features(df)
|
325 |
+
|
326 |
+
# Now gene_sequence is guaranteed to be a string, suitable for Seq
|
327 |
+
seq_obj = Seq(gene_sequence)
|
328 |
+
record = SeqRecord(seq_obj, id=gene_symbol, name=gene_symbol,
|
329 |
+
description=f'CRISPR Cas12 predicted targets for {gene_symbol}', features=features)
|
330 |
+
record.annotations["molecule_type"] = "DNA"
|
331 |
+
SeqIO.write(record, output_path, "genbank")
|
332 |
+
|
333 |
+
def create_bed_file_from_df(df, output_path):
|
334 |
+
with open(output_path, 'w') as bed_file:
|
335 |
+
for index, row in df.iterrows():
|
336 |
+
chrom = row["Chr"]
|
337 |
+
start = int(row["Target Start"])
|
338 |
+
end = start + len(row["Target"]) # Calculate the end position based on the target sequence length
|
339 |
+
strand = '+' if row["Strand"] == '1' else '-'
|
340 |
+
gRNA = row["gRNA"]
|
341 |
+
score = str(row["Prediction"])
|
342 |
+
is_mutation = 'Yes' if row["Is Mutation"] else 'No'
|
343 |
+
# transcript_id is not typically part of the standard BED columns but added here for completeness
|
344 |
+
transcript_id = row["Transcript"]
|
345 |
+
|
346 |
+
# Writing only standard BED columns; additional columns can be appended as needed
|
347 |
+
bed_file.write(f"{chrom}\t{start}\t{end}\t{gRNA}\t{score}\t{strand}\t{is_mutation}\n")
|
348 |
+
|
349 |
+
def create_csv_from_df(df, output_path):
|
350 |
+
df.to_csv(output_path, index=False)
|
351 |
|