File size: 8,072 Bytes
4f98c07
 
 
 
 
 
 
f2302fc
 
 
0a13e0b
f2302fc
0a13e0b
14e95ac
0a13e0b
4f98c07
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0a13e0b
67abe4f
4f98c07
f2302fc
67abe4f
aefd0c0
 
 
67abe4f
 
f2302fc
4f98c07
f2302fc
4f98c07
0a13e0b
f2302fc
 
4f98c07
f2302fc
 
0a13e0b
 
f2302fc
4f98c07
 
 
 
0a13e0b
4f98c07
 
 
 
 
 
 
 
 
 
f2302fc
629f8df
4f98c07
0a13e0b
 
f2302fc
0a13e0b
 
4f98c07
0a13e0b
 
 
 
 
 
4f98c07
0a13e0b
06b2426
4f98c07
0a13e0b
67abe4f
aefd0c0
 
f2302fc
 
06b2426
 
4f98c07
67abe4f
4f98c07
f2302fc
4f98c07
0a13e0b
 
4f98c07
06b2426
4f98c07
0a13e0b
4f98c07
f2302fc
06b2426
f2302fc
0a13e0b
 
67abe4f
 
06b2426
67abe4f
 
4f98c07
67abe4f
06b2426
 
 
67abe4f
 
4f98c07
06b2426
 
 
4f98c07
67abe4f
f2302fc
4f98c07
f2302fc
0a13e0b
06b2426
4f98c07
0a13e0b
4f98c07
 
 
0a13e0b
 
4f98c07
06b2426
4f98c07
0a13e0b
4f98c07
f2302fc
06b2426
f2302fc
0a13e0b
f2302fc
06b2426
0a13e0b
4f98c07
 
0a13e0b
4f98c07
 
 
 
f2302fc
 
4f98c07
06b2426
4f98c07
f2302fc
06b2426
 
 
 
 
 
 
f2302fc
0a13e0b
06b2426
0a13e0b
4f98c07
 
f2302fc
4f98c07
06b2426
 
 
 
97fec9a
06b2426
 
 
 
 
 
 
c9c4aaf
0a13e0b
06b2426
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f2302fc
06b2426
 
 
 
 
 
0a13e0b
06b2426
 
 
a1db08b
06b2426
 
 
 
 
 
 
 
 
 
f2302fc
4f98c07
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
import os
import json
import traceback
import joblib

import numpy as np
import pandas as pd

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt

import shap
import gradio as gr

# ============================================================
# PATCH: Fix gradio_client boolean JSON-schema crash
# ============================================================
try:
    import gradio_client.utils as _gcu

    if hasattr(_gcu, "_json_schema_to_python_type"):
        _orig = _gcu._json_schema_to_python_type

        def _patched(schema, defs=None):
            # JSON Schema allows boolean schemas (True / False)
            if isinstance(schema, bool):
                return "Any"
            return _orig(schema, defs)

        _gcu._json_schema_to_python_type = _patched
except Exception:
    pass


# ============================================================
# Artifacts
# ============================================================
MODEL_PATH = "model_extratrees_ds4.joblib"
FEATS_PATH = "feature_names.json"

model = joblib.load(MODEL_PATH)

with open(FEATS_PATH) as f:
    FEATURES = json.load(f)

LABEL_NAME = {0: "Benign", 1: "Pathogenic"}
THRESHOLD = 0.7

ID_CANDIDATES = [
    "#Uploaded_Variation", "#Uploaded_variation", "uploaded_variation",
    "Uploaded_Variation", "Uploaded_variation", "variant_id",
    "Variant_ID", "variation_id", "Variant", "ID", "HGVSc", "HGVSp"
]


# ============================================================
# Helpers
# ============================================================
def _clean_headers(df):
    df.columns = (
        df.columns.astype(str)
        .str.replace(r"[\u200b\ufeff]", "", regex=True)
        .str.strip()
    )
    return df


def _read_table(file):
    name = getattr(file, "name", str(file))
    if name.lower().endswith((".xlsx", ".xls")):
        df = pd.read_excel(name, engine="openpyxl")
    else:
        df = pd.read_csv(
            name,
            engine="python",
            sep=None,
            dtype=str,
            na_values=["", "NA", "N/A"],
            quoting=3,
            on_bad_lines="skip",
            encoding_errors="ignore",
        )
    return _clean_headers(df)


def _coerce_numeric(df):
    for c in df.columns:
        df[c] = pd.to_numeric(df[c], errors="coerce")
    return df


def _prepare(df):
    for c in FEATURES:
        if c not in df.columns:
            df[c] = np.nan
    return _coerce_numeric(df[FEATURES].copy())


def _validate_columns(df):
    return [c for c in FEATURES if c not in df.columns]


def _pick_id_series(df):
    norm = {c.lower(): c for c in df.columns}
    for cand in ID_CANDIDATES:
        if cand.lower() in norm:
            col = norm[cand.lower()]
            return df[col].astype(str)
    return pd.Series(df.index, name="row_index")


# ============================================================
# Prediction
# ============================================================
def predict_csv(file):
    try:
        if file is None:
            return None, "โŒ No file uploaded."

        df = _read_table(file)
        missing = _validate_columns(df)
        if missing:
            return None, f"โŒ Missing {len(missing)} features. Example: {missing[:10]}"

        X = _prepare(df)
        proba = model.predict_proba(X)[:, 1]
        pred = (proba > THRESHOLD).astype(int)

        ids = _pick_id_series(df)

        out = pd.DataFrame({
            "Variant ID": ids,
            "Predicted probability (pathogenic)": proba,
            "Predicted class": np.where(
                pred == 1, "๐Ÿ”ด Pathogenic", "๐ŸŸข Benign"
            ),
        })

        summary = (
            f"โœ… Processed {len(out)} variants  \n"
            f"๐Ÿ”ด Pathogenic: {(pred == 1).sum()}  |  "
            f"๐ŸŸข Benign: {(pred == 0).sum()}  \n"
            f"Threshold = {THRESHOLD}"
        )

        return out, summary

    except Exception:
        return None, "โŒ ERROR:\n```" + traceback.format_exc() + "```"


# ============================================================
# SHAP Explanation
# ============================================================
def explain_first_row(file):
    try:
        if file is None:
            return None, "โŒ No file uploaded."

        df = _read_table(file)
        missing = _validate_columns(df)
        if missing:
            return None, "โŒ Missing required features for explanation."

        X = _prepare(df)
        if len(X) == 0:
            return None, "โŒ No rows found."

        explainer = shap.Explainer(model.predict_proba, X)
        shap_values = explainer(X.iloc[[0]])

        vals = shap_values.values[0, :, 1]
        s = pd.Series(vals, index=FEATURES).sort_values(
            key=np.abs, ascending=False
        ).head(15)

        fig = plt.figure(figsize=(7, 5))
        s.iloc[::-1].plot(kind="barh")
        plt.title("Top feature contributions (Sample #1)")
        plt.tight_layout()

        return fig, (
            "๐Ÿ” **Why this prediction?**  \n"
            "Positive bars increase the probability of pathogenicity; "
            "negative bars decrease it.  \n"
            "These signals are *supportive only* and must be interpreted "
            "within ACMG/AMP guidelines."
        )

    except Exception:
        return None, "โŒ ERROR:\n```" + traceback.format_exc() + "```"


# ============================================================
# UI
# ============================================================
with gr.Blocks(
    title="Pathogenicity Predictor",
    theme=gr.themes.Soft(primary_hue="rose", secondary_hue="slate"),
) as demo:

    gr.Markdown("""
    # ๐Ÿงฌ Pathogenicity Predictor
    **Research prototype for variant pathogenicity assessment**

    Upload a CSV or Excel file containing the trained feature columns.
    This tool provides *supportive* predictions and explanations.
    """)

    with gr.Row():
        with gr.Column(scale=1):
            gr.Markdown("### ๐Ÿ“‚ Input")
            inp = gr.File(
                label="Upload CSV / XLSX",
                file_types=[".csv", ".xlsx", ".xls"]
            )

            gr.Markdown("""
            **Requirements**
            - One row = one variant  
            - Must include all trained feature columns  
            - Results are **not** clinical diagnoses
            """)

        with gr.Column(scale=2):
            gr.Markdown("### ๐Ÿ“Š Results")

            with gr.Row():
                btn_pred = gr.Button("๐Ÿ”ฎ Predict", variant="primary")
                btn_expl = gr.Button("๐Ÿ” Explain first variant")

            out_df = gr.Dataframe(
                label="Predictions",
                interactive=False,
                wrap=True,
            )

            out_txt = gr.Markdown()

    gr.Markdown("### ๐Ÿ”ฌ Why this prediction?")
    out_plot = gr.Plot(label="Feature contribution (SHAP)")

    btn_pred.click(predict_csv, inp, [out_df, out_txt], show_progress=True)
    btn_expl.click(explain_first_row, inp, [out_plot, out_txt])

    # ========================================================
    # Feedback section (email)
    # ========================================================
    gr.Markdown("""
    ---
    ### ๐Ÿ’ฌ Feedback

    We would love to hear your thoughts, suggestions, or bug reports.

    ๐Ÿ“ง **[Send feedback via email](mailto:dr.rahafahmad@gmail.com?subject=Pathogenicity%20Predictor%20Feedback)**
    Click on the link or email us directly at **dr.rahafahmad@gmail.com, 700039113@uaeu.ac.ae, bassam.ali@uaeu.ac.ae**
    """)

    gr.Markdown("""
    ---
    โš ๏ธ **Research prototype**  
    This tool is for research and educational use only.  
    Not intended for clinical diagnosis or medical decision-making.

    ยฉ 2026
    """)


# ============================================================
# Launch (Spaces-safe)
# ============================================================
if __name__ == "__main__":
    port = int(os.environ.get("PORT", 7860))
    demo.queue()
    demo.launch(server_name="0.0.0.0", server_port=port, debug=True)