dev: first commit

README.md

@@ -1,14 +1,136 @@
 ---
-title: ChimeraLM
-emoji: 🔥
-colorFrom: purple
+title: ChimeraLM - Chimeric Read Detector
+emoji: 🧬
+colorFrom: blue
 colorTo: purple
 sdk: gradio
-sdk_version: 5.49.1
+sdk_version: "5.0.0"
 app_file: app.py
-pinned: false
+pinned: true
 license: apache-2.0
-short_description: 'A genomic language model that distinguishes true structural '
+tags:
+  - genomics
+  - bioinformatics
+  - deep-learning
+  - dna-sequence
+  - chimera-detection
+  - whole-genome-amplification
+  - pytorch
+  - lightning
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# 🧬 ChimeraLM: Chimeric Read Detector
+
+<div align="center">
+
+![ChimeraLM](https://img.shields.io/badge/ChimeraLM-v1.0.4-blue?style=for-the-badge)
+![Python](https://img.shields.io/badge/Python-3.10+-green?style=for-the-badge)
+![PyTorch](https://img.shields.io/badge/PyTorch-2.5-orange?style=for-the-badge)
+![License](https://img.shields.io/badge/License-Apache_2.0-red?style=for-the-badge)
+
+**Advanced Chimeric Read Detection using Deep Learning**
+
+[🏠 Homepage](https://github.com/ylab-hi/ChimeraLM) | [📚 Documentation](https://ylab-hi.github.io/ChimeraLM/) | [🤗 Model](https://huggingface.co/yangliz5/chimeralm) | [📦 PyPI](https://pypi.org/project/chimeralm/)
+
+</div>
+
+---
+
+## 🚀 What is ChimeraLM?
+
+ChimeraLM is a state-of-the-art genomic language model designed to identify **chimeric artifacts** introduced by whole genome amplification (WGA). Chimeric reads are artificial DNA sequences where fragments from different genomic locations are incorrectly joined together during the amplification process.
+
+### ⚡ Key Features
+
+- **🎯 High Accuracy**: 98%+ accuracy in detecting chimeric vs biological reads
+- **⚡ Fast Inference**: Optimized for both CPU and GPU (CUDA/MPS)
+- **📏 Long Sequences**: Supports DNA sequences up to 32,768 nucleotides
+- **🤖 Pre-trained Model**: Ready-to-use model from Hugging Face Hub
+- **🔬 Research-Grade**: Trained on real WGA data from genomic studies
+
+### 🧬 How It Works
+
+1. **Input**: DNA sequence (A, C, G, T, N nucleotides)
+2. **Processing**: HyenaDNA-based transformer model analyzes the sequence
+3. **Output**: Binary classification (Biological vs Chimeric) with confidence scores
+
+---
+
+## 💡 Use Cases
+
+- **Quality Control**: Filter chimeric artifacts from WGA sequencing data
+- **Genomic Analysis**: Improve accuracy of variant calling and assembly
+- **Research**: Study patterns in whole genome amplification
+- **Education**: Learn about chimeric artifacts and deep learning in genomics
+
+---
+
+## 🛠️ Installation & CLI Usage
+
+For batch processing and production use, install the CLI tool:
+
+```bash
+# Install via pip
+pip install chimeralm
+
+# Predict chimeric reads
+chimeralm predict your_data.bam --gpus 1 --batch-size 24
+
+# Filter BAM to remove chimeric reads
+chimeralm filter your_data.bam predictions/
+```
+
+**Requirements**: Python 3.10, 3.11, or 3.12
+
+---
+
+## 📊 Model Architecture
+
+- **Backbone**: HyenaDNA-small-32k (256-dim embeddings)
+- **Head**: Binary Sequence Classifier with attention pooling
+- **Loss**: CrossEntropyLoss
+- **Optimizer**: AdamW (lr=1e-4, weight_decay=0.01)
+- **Training Data**: Real WGA chimeric artifacts from genomic studies
+
+**Model Size**: ~50M parameters
+**Inference Speed**: ~1000 sequences/second (GPU)
+
+---
+
+## 📖 Citation
+
+If you use ChimeraLM in your research, please cite:
+
+```bibtex
+@software{chimeralm2025,
+  title={ChimeraLM: A genomic language model to identify chimera artifacts},
+  author={Li, Yangyang and Guo, Qingxiang and Yang, Rendong},
+  year={2025},
+  url={https://github.com/ylab-hi/ChimeraLM}
+}
+```
+
+---
+
+## 🔗 Resources
+
+- **GitHub**: [ylab-hi/ChimeraLM](https://github.com/ylab-hi/ChimeraLM)
+- **Documentation**: [ylab-hi.github.io/ChimeraLM](https://ylab-hi.github.io/ChimeraLM/)
+- **Model Hub**: [yangliz5/chimeralm](https://huggingface.co/yangliz5/chimeralm)
+- **PyPI Package**: [pypi.org/project/chimeralm](https://pypi.org/project/chimeralm/)
+
+---
+
+## 📝 License
+
+This project is licensed under the **Apache License 2.0** - see the [LICENSE](https://github.com/ylab-hi/ChimeraLM/blob/main/LICENSE) file for details.
+
+---
+
+<div align="center">
+
+**⭐ Star us on GitHub if you find this useful!**
+
+[⭐ Star on GitHub](https://github.com/ylab-hi/ChimeraLM) | [🐛 Report Bug](https://github.com/ylab-hi/ChimeraLM/issues) | [💡 Request Feature](https://github.com/ylab-hi/ChimeraLM/issues)
+
+</div>

app.py

@@ -0,0 +1,769 @@
+"""Gradio Web UI for ChimeraLM - Hugging Face Spaces Version."""
+
+import logging
+import os
+
+import gradio as gr
+import plotly.graph_objects as go
+import torch
+
+import chimeralm
+from chimeralm.data.tokenizer import load_tokenizer_from_hyena_model
+
+# Set up logging
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
+)
+logger = logging.getLogger(__name__)
+
+
+class ChimeraLMPredictor:
+    """ChimeraLM predictor for web interface."""
+
+    def __init__(self):
+        self.model = None
+        self.tokenizer = None
+        self.device = "cuda" if torch.cuda.is_available() else "cpu"
+        logger.info(f"Using device: {self.device}")
+        self._load_model()
+
+    def _load_model(self):
+        """Load the ChimeraLM model and tokenizer."""
+        try:
+            logger.info("Loading ChimeraLM model from Hugging Face Hub...")
+            self.model = chimeralm.models.ChimeraLM.from_pretrained("yangliz5/chimeralm")
+            self.model.eval()
+            self.model.to(self.device)
+
+            logger.info("Loading tokenizer...")
+            self.tokenizer = load_tokenizer_from_hyena_model("hyenadna-small-32k-seqlen")
+            logger.info(f"✅ Model loaded successfully on {self.device}")
+        except Exception as e:
+            logger.error(f"❌ Failed to load model: {e}")
+            raise
+
+    def predict(self, sequence: str) -> tuple[str, float, dict]:
+        """Predict if a DNA sequence is chimeric or biological."""
+        if not sequence or not sequence.strip():
+            return "Please enter a DNA sequence", 0.0, {}
+
+        # Clean and validate sequence
+        sequence = sequence.strip().upper()
+        valid_chars = set("ACGTNacgtn")
+        if not all(c in valid_chars for c in sequence):
+            return "Invalid characters in sequence. Only A, C, G, T, N are allowed.", 0.0, {}
+
+        sequence = sequence.upper()
+
+        try:
+            # Tokenize sequence
+            tokenized = self.tokenizer(
+                sequence,
+                truncation=True,
+                padding=True,
+                max_length=32768,
+                return_tensors="pt",
+            )
+
+            # Extract input_ids and move to device
+            input_ids = tokenized["input_ids"].to(self.device)
+            input_quals = None  # We don't have quality scores for web input
+
+            # Make prediction
+            with torch.no_grad():
+                logits = self.model(input_ids, input_quals)
+                probabilities = torch.softmax(logits, dim=-1)
+                predicted_class = torch.argmax(probabilities, dim=-1).item()
+                confidence = probabilities[0][predicted_class].item()
+
+            # Interpret results
+            class_names = ["Biological", "Chimeric Artifact"]
+            prediction = class_names[predicted_class]
+
+            # Create confidence breakdown
+            confidence_breakdown = {
+                "Biological": f"{probabilities[0][0].item():.3f}",
+                "Chimeric Artifact": f"{probabilities[0][1].item():.3f}",
+            }
+
+            logger.info(f"Prediction: {prediction} (confidence: {confidence:.3f})")
+            return prediction, confidence, confidence_breakdown
+
+        except Exception as e:
+            logger.error(f"Prediction error: {e}")
+            return f"Prediction failed: {e}", 0.0, {}
+
+
+def create_interface():
+    """Create the Gradio interface."""
+    predictor = ChimeraLMPredictor()
+
+    def predict_sequence(sequence):
+        prediction, confidence, breakdown = predictor.predict(sequence)
+
+        # Format output with enhanced styling
+        if "❌" in prediction or "⚠️" in prediction or "Please" in prediction or "Invalid" in prediction or "Prediction failed" in prediction:
+            result_text = f"### {prediction}"
+        else:
+            # Color-coded results with better styling
+            color = "#4CAF50" if prediction == "Biological" else "#F44336"
+            icon = "✅" if prediction == "Biological" else "⚠️"
+            result_text = f"""
+### {icon} Prediction Result
+
+<div style="background: {color}; color: white; padding: 1.5rem; border-radius: 15px; text-align: center; margin: 1rem 0; box-shadow: 0 4px 15px rgba(0,0,0,0.15);">
+    <h2 style="margin: 0; font-size: 2rem; font-weight: 700; color: white;">{prediction}</h2>
+    <p style="margin: 0.5rem 0 0 0; font-size: 1.2rem; color: rgba(255,255,255,0.95);">Confidence: {confidence:.1%}</p>
+</div>
+"""
+
+        if breakdown:
+            result_text += "\n\n### 📊 Detailed Confidence Scores:\n"
+            for class_name, prob in breakdown.items():
+                emoji = "✅" if class_name == "Biological" else "⚠️"
+                prob_value = float(prob)
+                result_text += f"- {emoji} **{class_name}**: {prob_value:.1%}\n"
+
+        # Create bar plot with proper contrast
+        if breakdown:
+            classes = list(breakdown.keys())
+            probabilities = [float(prob) for prob in breakdown.values()]
+
+            # Create colors based on prediction with better contrast
+            colors = []
+            text_colors = []
+            for class_name in classes:
+                if class_name == prediction:
+                    # Vibrant colors for predicted class with white text
+                    if prediction == "Biological":
+                        colors.append("#4CAF50")  # Green
+                    else:
+                        colors.append("#F44336")  # Red
+                    text_colors.append("white")
+                else:
+                    # Medium gray for non-predicted class with dark text
+                    colors.append("#BDBDBD")
+                    text_colors.append("#424242")
+
+            # Create individual bars with appropriate text colors
+            bars = []
+            for i, (class_name, prob, color, text_color) in enumerate(zip(classes, probabilities, colors, text_colors)):
+                bars.append(
+                    go.Bar(
+                        x=[class_name],
+                        y=[prob],
+                        marker_color=color,
+                        text=[f"{prob:.1%}"],
+                        textposition="auto",
+                        textfont={"size": 20, "color": text_color, "family": "Inter, sans-serif", "weight": 600},
+                        marker_line={"width": 2, "color": "rgba(255,255,255,0.3)"},
+                        width=0.5,
+                        opacity=0.95,
+                        name=class_name,
+                        showlegend=False,
+                    )
+                )
+
+            fig = go.Figure(data=bars)
+
+            fig.update_layout(
+                title={
+                    "text": "🎯 Prediction Confidence",
+                    "font": {"size": 20, "color": "#424242", "family": "Arial, sans-serif"},
+                    "x": 0.5,
+                    "xanchor": "center",
+                },
+                xaxis={
+                    "title": {"text": "Classification", "font": {"size": 14, "color": "#616161"}},
+                    "tickfont": {"size": 12, "color": "#424242"},
+                    "gridcolor": "rgba(0,0,0,0.05)",
+                    "linecolor": "rgba(0,0,0,0.1)",
+                    "showgrid": True,
+                    "zeroline": False,
+                },
+                yaxis={
+                    "title": {"text": "Probability", "font": {"size": 14, "color": "#616161"}},
+                    "tickfont": {"size": 12, "color": "#424242"},
+                    "range": [0, 1.1],
+                    "gridcolor": "rgba(0,0,0,0.05)",
+                    "linecolor": "rgba(0,0,0,0.1)",
+                    "showgrid": True,
+                    "zeroline": True,
+                    "zerolinecolor": "rgba(0,0,0,0.1)",
+                },
+                height=450,
+                showlegend=False,
+                plot_bgcolor="rgba(255,255,255,1)",
+                paper_bgcolor="rgba(255,255,255,1)",
+                margin={"l": 60, "r": 60, "t": 80, "b": 60},
+                font={"family": "Arial, sans-serif"},
+            )
+
+            fig.update_traces(
+                textfont_size=16,
+                textfont_color="white",
+                textfont_family="Arial, sans-serif",
+                marker_line={"width": 1, "color": "rgba(255,255,255,0.8)"},
+                width=0.6,
+                opacity=0.9,
+            )
+        else:
+            # Create empty plot for error cases
+            fig = go.Figure()
+            fig.update_layout(
+                title={
+                    "text": "🎯 Prediction Confidence",
+                    "font": {"size": 20, "color": "#424242", "family": "Arial, sans-serif"},
+                    "x": 0.5,
+                    "xanchor": "center",
+                },
+                xaxis={
+                    "title": {"text": "Classification", "font": {"size": 14, "color": "#616161"}},
+                    "tickfont": {"size": 12, "color": "#424242"},
+                    "gridcolor": "rgba(0,0,0,0.05)",
+                    "linecolor": "rgba(0,0,0,0.1)",
+                },
+                yaxis={
+                    "title": {"text": "Probability", "font": {"size": 14, "color": "#616161"}},
+                    "tickfont": {"size": 12, "color": "#424242"},
+                    "range": [0, 1.1],
+                    "gridcolor": "rgba(0,0,0,0.05)",
+                    "linecolor": "rgba(0,0,0,0.1)",
+                },
+                height=450,
+                showlegend=False,
+                plot_bgcolor="rgba(255,255,255,1)",
+                paper_bgcolor="rgba(255,255,255,1)",
+                margin={"l": 60, "r": 60, "t": 80, "b": 60},
+                font={"family": "Arial, sans-serif"},
+            )
+
+        return result_text, fig
+
+    # Example sequences - more realistic with varied patterns
+    # 1, 1, 0
+    examples = [
+        ["TTGTGTGCCTTCATTAGTTATATACTAGTTCCTGATAAATTCATTTATAGAACAGAAAGACCACAGATTCAATTATATGGAATAGATCTGCTGGTGAATGTAAGAAAGTCTTCTGAACTGCGAAGGGAAAATAAATGATTTAATTCCCACCACCTCTCAACAGCTACCTTCTGTTTTAGAGACACTGGTAAAACTTCTGGGGCTCTTACTTGACATACCTACATCGTATTATAGGCCTATTGGTTTTATCAGAATAATATGCTTTCCTCACATAAGTTATTTCTTTCTGTTACTTGCTTGCAGTACAGATTTAAAGGGGCATTCAGGCAGCCTCCAGATGCCATGATGGATTAACTCTCATGTTACACAGTAATGTAGAAGCTTCTCTTCATTCTCAGACTTTATCTGACAATGAAGAGAAGCTTCTAATTATACTGTGTAAGTTGATCATGTAACACATCTGGAGGCTGCCTGAATGCCCCTTAAATCTGTACTGCAAGCAAGTAACAGAAAGAAATAACTTATGTGAGGAAAGCATATTATTCTGATAAAACCAATACACCCTTATAATACGATGTAGGTATGTCAAGTAAGAGCCCCAGAAGTTTGCAGTATCTAAAACAAAGGTGTTTGTTGAGGTAGTGAGGAAAATAAATCATTTATTTTCCCTTCGCAGTTCAGGCAACACTTTCTTTACATTCACCACCAGATTCCATATAATCTGTGGGAGTCTTTGGCTGTTCTATAAAATGAATTTATCAGTAAATGA"],
+        ["CAATGGTAAATGAATTCAATAAATATTTGAGGTGATTAAATTTCCTTTCCTAACACATTTTATTTCAAATTCTATTTGAAAGAAAAAATGCTAACAACATAAGAGATCAAATTCAGCTACCTATTTTTTCAACATTCAAATATGCATTAATTGTCTACACTTTGCTAAGCTTGGGCTGATTTCTAGGGCTATAAACATAAATTAAATTTATTCATGGATCTTAAGTGGCTCATGAGCATTAGTACAGCATATTTATAAGCCGAGCATAGTGTCTCATACCTATAATCCCAACACTGGGAGGCTGAGGTGGGAGGATCTCTTGAAGCCGGGAGTTCAAGAACTGCCTGGAAAACATAGCAAGACCCTGTCTCTACCAAAAACAAACAAATAAAACTTAGCCGGGAGTGGCTGCACCTGTAGCTACTCAGGAGTCTGTGATTGGAGGGTAATTTGAACACAGGAGTTTGAGATAGCAGCAAGCTATGATCATGCCACTGTACTCCAGCCTAATTGACAGAACAAGAGCCTGTCTCTAAAATCATTCCATATGTCTATATATAGATATATATATCAAGAAAACTTTACTTTCTAGATTCTAGTTTGTTTTATTGCTCATTCTTTTCTAAATTTATTCATTAGGAGGTATATACAATGTGTTTCAGAGATATAAGAATAGTAAACTTAGAGTGAAAAGGGAAAGATATTTCTTGTTAAAATTCCTAAAATAAAGTATTAAACTTATCTATGAAAAGGCATACATTTCTGTCTGATATTTTATATAAAATAATGGGAACATAATCATATATAATATTTTCTATAAAATGCTTAACAGGTTTTCATAACTTAAATTGTACTTAATATTTTAGGAATTTTAACAATATTCTTCCCTTTTCACTCTAAGTTTACTGTCTTAACCCCCAAAAAACACATTGTCTGTACACCTCCTAATGAATAAATTTAGAAAAAGAAAAAATACAGCAATAAAACAAACTAGTAATACTGGAAGAGTCAAACTTTCTGATATTGTGTACCTCTTCTTATAAAGACATATGGAATGATTTTGAGGACAGGTATTGTTCTGATTAGGCTGGAGTACAGTGGCATGATCATAGCTTACTGCTATCTCGAACTCCTGTGTTAAATTCTCTCCAATCACAGACTCCTGAGTAGCTACAGGTGAGCCACTGCCCGGCTAAGTTTTATTTGTTTTGTTTTTGGTAGAGACAGGGTCTTGCTATGTTTGCCAGGCTGGTCTTGAACTCCCGGCTTCAAGAGATCCTCCCACCTCAGCCTCCCAGTGTTGGGATTATAGGTATGAGACACTATGCTCAGCTAACAAATATATAATGCTCATGAGCCACTAATCAAGTCAAGAATTTAAATTTATGTTTATAGCCCCATCAGCCCCAAGCTTAGCAAAGTGTAGACAATTAATGTAACATTTGAATGCTGAAAAAATAGGTATAGAAATTTGATCTTACCCTATATTGTAGCATTTTTTTCTTTCAAATAAATTTGAAATAAAATGTGTAGGGAAAAGGAAATTAAATCACCTCAACATTTTATAAAAATCATTTACCATTGGCTAT"],
+        ["ATGTTGTGTACCTGGTTCGGTTCGTCTATGGTATGCACCTTGGCTATCATCACCCGATGAGGCAACCAGCCGGGAGACACCTAAACCCATCATCTCCTGTACCACCCTAGTAGGCTCCCTTCCCCTACTCATCGCACTAATTTACACTCACAACACCCTAGGCTCACTAAACATTCTACTACTCACTCTCACTGCCCAACTAAACTCCTGGCCATCCCCTTATGAGCGGGCGCAGTGATTATAGGCTTTCGCTCTAAGATTAAAAATGCCCTAGCCCACTTCTTACCACAAGGCACACCTACACCCCTTATCCCCATACTGGCTGTTGTGAAAACCATAGCCTACTATCGTTCAACAATAGCCCTGGCCGTACGCCTAACCGCTAACATTACTGCAGGCCACCTACTCATGCACCTAATTGGAAGCGCCACCCTAGCAATATCAACCATTAACCTTACCTACACTTATAGTCTTTCACAATTCTAATTCTACTGACTATCCTAGAAATCGCTGTCGCCTTAATCCAAGCCTACGTTTTCACACTTCTAGTAAGCCTCTACCTGCACGACAACACATAATGACCCACCAATCACATGCCTATCATGGCTAAACCCAGCCCATGACCCCTAACAGGGGCCCTCTCAGCCCTCCTAATGACCTCCGGCCTAGCCATGTGATTTCACTTCCACTCCATAACGCTCCTCATACTAGGCCTACTAACCAACACACTAACCATATACCAATAATGGCAATGTAACGCAAAGCACATACCAAGGCCACCACACACCACCTCTATTAAAAAGGCC"],
+    ]
+
+    # Custom CSS for modern, visually appealing styling
+    custom_css = """
+    @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');
+
+    * {
+        font-family: 'Inter', -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif;
+    }
+
+    /* Global text color improvements */
+    body {
+        background: linear-gradient(135deg, #f5f7fa 0%, #e3e7ed 100%);
+        min-height: 100vh;
+    }
+
+    /* Ensure all headings have good contrast */
+    h1, h2, h3, h4, h5, h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    /* Ensure all paragraphs and text have good contrast */
+    p, li, span, div {
+        color: #37474F !important;
+    }
+
+    /* Universal text color fix for all content */
+    strong, b {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    /* Ensure all text in Gradio blocks has proper contrast */
+    .gradio-block p, .gradio-block li, .gradio-block span,
+    .gradio-block div, .gradio-block strong, .gradio-block b {
+        color: #37474F !important;
+    }
+
+    .gradio-block h1, .gradio-block h2, .gradio-block h3,
+    .gradio-block h4, .gradio-block h5, .gradio-block h6,
+    .gradio-block strong, .gradio-block b {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    /* Label styling */
+    label {
+        color: #2C3E50 !important;
+        font-weight: 600 !important;
+    }
+
+    .main-header {
+        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+        color: white;
+        padding: 3rem 2rem;
+        border-radius: 20px;
+        text-align: center;
+        margin-bottom: 2rem;
+        box-shadow: 0 15px 40px rgba(102, 126, 234, 0.3);
+        position: relative;
+        overflow: hidden;
+    }
+
+    .main-header::before {
+        content: '';
+        position: absolute;
+        top: -50%;
+        right: -50%;
+        bottom: -50%;
+        left: -50%;
+        background: linear-gradient(45deg, transparent, rgba(255,255,255,0.1), transparent);
+        transform: rotate(45deg);
+        animation: shine 3s infinite;
+    }
+
+    @keyframes shine {
+        0% { transform: translateX(-100%) rotate(45deg); }
+        100% { transform: translateX(100%) rotate(45deg); }
+    }
+
+    .dna-icon {
+        font-size: 4rem;
+        margin-bottom: 1rem;
+        animation: pulse 2s ease-in-out infinite;
+        display: inline-block;
+        filter: drop-shadow(0 4px 6px rgba(0,0,0,0.2));
+    }
+
+    @keyframes pulse {
+        0%, 100% { transform: scale(1); }
+        50% { transform: scale(1.08); }
+    }
+
+    .input-column {
+        background: white;
+        padding: 2.5rem;
+        border-radius: 20px;
+        box-shadow: 0 8px 30px rgba(0,0,0,0.1);
+        margin: 0.5rem;
+        border: 1px solid rgba(102, 126, 234, 0.1);
+        transition: transform 0.3s ease, box-shadow 0.3s ease;
+    }
+
+    .input-column:hover {
+        transform: translateY(-2px);
+        box-shadow: 0 12px 40px rgba(0,0,0,0.15);
+    }
+
+    /* Ensure input column text has good contrast */
+    .input-column h1, .input-column h2, .input-column h3,
+    .input-column h4, .input-column h5, .input-column h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .input-column p, .input-column li, .input-column span,
+    .input-column div, .input-column strong, .input-column b,
+    .input-column code, .input-column pre {
+        color: #37474F !important;
+    }
+
+    /* Ensure markdown content in input column has proper colors */
+    .input-column .markdown, .input-column .markdown *,
+    .input-column [class*="markdown"], .input-column [class*="markdown"] * {
+        color: #37474F !important;
+    }
+
+    .input-column .markdown h1, .input-column .markdown h2, .input-column .markdown h3,
+    .input-column [class*="markdown"] h1, .input-column [class*="markdown"] h2, .input-column [class*="markdown"] h3 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .input-column .markdown strong, .input-column .markdown b,
+    .input-column [class*="markdown"] strong, .input-column [class*="markdown"] b {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .result-column {
+        background: linear-gradient(135deg, #ffffff 0%, #f8f9fa 100%);
+        padding: 2.5rem;
+        border-radius: 20px;
+        box-shadow: 0 8px 30px rgba(0,0,0,0.1);
+        margin: 0.5rem;
+        border: 1px solid rgba(102, 126, 234, 0.1);
+        min-height: 500px;
+    }
+
+    /* Ensure text readability in result column */
+    .result-column h1, .result-column h2, .result-column h3,
+    .result-column h4, .result-column h5, .result-column h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .result-column p, .result-column li, .result-column span,
+    .result-column div, .result-column markdown {
+        color: #37474F !important;
+    }
+
+    /* Markdown content styling - comprehensive */
+    .markdown, .markdown *, [class*="markdown"], [class*="prose"] {
+        color: #37474F !important;
+    }
+
+    .markdown h1, .markdown h2, .markdown h3,
+    .markdown h4, .markdown h5, .markdown h6,
+    [class*="markdown"] h1, [class*="markdown"] h2, [class*="markdown"] h3,
+    [class*="markdown"] h4, [class*="markdown"] h5, [class*="markdown"] h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .markdown p, .markdown li, .markdown span,
+    .markdown div, .markdown code, .markdown pre,
+    .markdown strong, .markdown b,
+    [class*="markdown"] p, [class*="markdown"] li, [class*="markdown"] span,
+    [class*="markdown"] div, [class*="markdown"] strong, [class*="markdown"] b {
+        color: #37474F !important;
+    }
+
+    .markdown code, [class*="markdown"] code {
+        background: #f5f7fa !important;
+        color: #2C3E50 !important;
+        padding: 2px 6px !important;
+        border-radius: 4px !important;
+    }
+
+    /* Target all Gradio markdown blocks */
+    .gradio-markdown, .gradio-markdown *,
+    div[class*="markdown"], div[class*="prose"] {
+        color: #37474F !important;
+    }
+
+    div[class*="markdown"] h1, div[class*="markdown"] h2, div[class*="markdown"] h3,
+    div[class*="markdown"] h4, div[class*="markdown"] h5, div[class*="markdown"] h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    div[class*="markdown"] strong, div[class*="markdown"] b,
+    div[class*="markdown"] p, div[class*="markdown"] li,
+    div[class*="markdown"] span, div[class*="markdown"] div {
+        color: #37474F !important;
+    }
+
+    .footer-section {
+        background: linear-gradient(135deg, #ffffff 0%, #f8f9fa 100%);
+        padding: 2.5rem;
+        border-radius: 20px;
+        margin-top: 2rem;
+        border: 2px solid #dee2e6;
+        box-shadow: 0 5px 20px rgba(0,0,0,0.08);
+    }
+
+    /* Ensure footer text has good contrast */
+    .footer-section h1, .footer-section h2, .footer-section h3,
+    .footer-section h4, .footer-section h5, .footer-section h6 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .footer-section p, .footer-section li, .footer-section span,
+    .footer-section div, .footer-section a, .footer-section code,
+    .footer-section strong, .footer-section b {
+        color: #37474F !important;
+    }
+
+    /* Ensure markdown content in footer has proper colors */
+    .footer-section .markdown, .footer-section .markdown *,
+    .footer-section [class*="markdown"], .footer-section [class*="markdown"] * {
+        color: #37474F !important;
+    }
+
+    .footer-section .markdown h1, .footer-section .markdown h2, .footer-section .markdown h3,
+    .footer-section [class*="markdown"] h1, .footer-section [class*="markdown"] h2, .footer-section [class*="markdown"] h3 {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .footer-section .markdown strong, .footer-section .markdown b,
+    .footer-section [class*="markdown"] strong, .footer-section [class*="markdown"] b {
+        color: #2C3E50 !important;
+        font-weight: 700 !important;
+    }
+
+    .footer-section a {
+        color: #667eea !important;
+        text-decoration: none !important;
+        font-weight: 600 !important;
+    }
+
+    .footer-section a:hover {
+        color: #764ba2 !important;
+        text-decoration: underline !important;
+    }
+
+    .footer-section code {
+        background: #f5f7fa !important;
+        color: #2C3E50 !important;
+        padding: 2px 6px !important;
+        border-radius: 4px !important;
+        border: 1px solid #e0e0e0 !important;
+    }
+
+    .gradio-container {
+        max-width: 1400px !important;
+        margin: 0 auto !important;
+        padding: 2rem 1rem !important;
+    }
+
+    .analyze-btn {
+        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%) !important;
+        border: none !important;
+        border-radius: 30px !important;
+        padding: 18px 40px !important;
+        font-size: 17px !important;
+        font-weight: 600 !important;
+        color: white !important;
+        box-shadow: 0 6px 20px rgba(102, 126, 234, 0.4) !important;
+        transition: all 0.3s cubic-bezier(0.4, 0, 0.2, 1) !important;
+        text-transform: uppercase;
+        letter-spacing: 0.5px;
+    }
+
+    .analyze-btn:hover {
+        transform: translateY(-3px) scale(1.02) !important;
+        box-shadow: 0 12px 35px rgba(102, 126, 234, 0.6) !important;
+    }
+
+    .analyze-btn:active {
+        transform: translateY(-1px) scale(0.98) !important;
+    }
+
+    /* Enhanced textbox styling */
+    textarea {
+        border: 2px solid #e9ecef !important;
+        border-radius: 12px !important;
+        font-family: 'Courier New', monospace !important;
+        font-size: 14px !important;
+        line-height: 1.6 !important;
+        transition: border-color 0.3s ease !important;
+        background-color: #ffffff !important;
+        color: #2C3E50 !important;
+        padding: 14px !important;
+    }
+
+    textarea::placeholder {
+        color: #90A4AE !important;
+        opacity: 0.8 !important;
+    }
+
+    textarea:focus {
+        border-color: #667eea !important;
+        box-shadow: 0 0 0 3px rgba(102, 126, 234, 0.1) !important;
+        background-color: #ffffff !important;
+        outline: none !important;
+    }
+
+    /* Info cards */
+    .info-card {
+        background: white;
+        padding: 1.5rem;
+        border-radius: 15px;
+        box-shadow: 0 4px 15px rgba(0,0,0,0.08);
+        margin: 1rem 0;
+        border-left: 4px solid #667eea;
+        transition: transform 0.2s ease;
+    }
+
+    .info-card:hover {
+        transform: translateX(5px);
+    }
+
+    /* Examples styling */
+    #examples {
+        border-radius: 15px;
+        overflow: hidden;
+        margin-top: 1.5rem;
+    }
+
+    /* Enhanced examples button styling */
+    .example-btn {
+        background: #f8f9fa !important;
+        border: 2px solid #e0e0e0 !important;
+        color: #2C3E50 !important;
+        border-radius: 8px !important;
+        padding: 12px 20px !important;
+        transition: all 0.3s ease !important;
+    }
+
+    .example-btn:hover {
+        background: #667eea !important;
+        border-color: #667eea !important;
+        color: white !important;
+        transform: translateY(-2px) !important;
+        box-shadow: 0 4px 12px rgba(102, 126, 234, 0.3) !important;
+    }
+
+    /* Better spacing and visual hierarchy */
+    .gradio-block {
+        margin-bottom: 1.5rem;
+    }
+
+    /* Improved scrollbar styling */
+    ::-webkit-scrollbar {
+        width: 10px;
+        height: 10px;
+    }
+
+    ::-webkit-scrollbar-track {
+        background: #f1f1f1;
+        border-radius: 10px;
+    }
+
+    ::-webkit-scrollbar-thumb {
+        background: #667eea;
+        border-radius: 10px;
+    }
+
+    ::-webkit-scrollbar-thumb:hover {
+        background: #764ba2;
+    }
+    """
+
+    with gr.Blocks(
+        title="ChimeraLM - Chimeric Read Detector",
+        theme=gr.themes.Default(
+            primary_hue="blue",
+            secondary_hue="gray",
+            neutral_hue="slate",
+        ),
+        css=custom_css,
+    ) as interface:
+        # Header Section
+        with gr.Row():
+            gr.HTML("""
+                <div class="main-header">
+                    <div class="dna-icon">🧬</div>
+                    <h1 style="margin: 0; font-size: 3rem; font-weight: 700; position: relative; z-index: 1;">ChimeraLM</h1>
+                    <p style="margin: 0.5rem 0 0 0; font-size: 1.3rem; opacity: 0.95; font-weight: 500; position: relative; z-index: 1;">
+                        Advanced Chimeric Read Detection using Deep Learning
+                    </p>
+                    <p style="margin: 1rem 0 0 0; font-size: 1.05rem; opacity: 0.85; position: relative; z-index: 1;">
+                        Identify chimeric artifacts from whole genome amplification with state-of-the-art accuracy
+                    </p>
+                    <div style="margin-top: 1.5rem; position: relative; z-index: 1;">
+                        <span style="display: inline-block; background: rgba(255,255,255,0.2); padding: 0.5rem 1rem; border-radius: 20px; margin: 0.25rem; font-size: 0.9rem;">
+                            ⚡ High Performance
+                        </span>
+                        <span style="display: inline-block; background: rgba(255,255,255,0.2); padding: 0.5rem 1rem; border-radius: 20px; margin: 0.25rem; font-size: 0.9rem;">
+                            🎯 98% Accuracy
+                        </span>
+                        <span style="display: inline-block; background: rgba(255,255,255,0.2); padding: 0.5rem 1rem; border-radius: 20px; margin: 0.25rem; font-size: 0.9rem;">
+                            🚀 Pre-trained
+                        </span>
+                    </div>
+                </div>
+            """)
+
+        # Main Content
+        with gr.Row():
+            with gr.Column(scale=1, elem_classes="input-column"):
+                # Input Section
+                gr.Markdown("""
+                ## 📝 DNA Sequence Input
+
+                **Quick Start Guide:**
+                1. 🧬 Enter your DNA sequence (supports up to 32,768 bp)
+                2. ✅ Use standard nucleotides: **A**, **C**, **G**, **T**, **N**
+                3. 🔬 Click "Analyze Sequence" for instant results
+                4. 📊 View confidence scores and visualization below
+
+                **What is Chimeric DNA?**
+                Chimeric reads are artificial DNA sequences created during whole genome amplification (WGA),
+                where fragments from different genomic locations are incorrectly joined together.
+                """)
+
+                sequence_input = gr.Textbox(
+                    label="🧬 DNA Sequence",
+                    placeholder="Enter your DNA sequence here...\nExample: ACGTACGTACGTACGT...",
+                    lines=8,
+                    max_lines=15,
+                    show_label=True,
+                    container=True,
+                    scale=2,
+                )
+
+                with gr.Row():
+                    predict_btn = gr.Button(
+                        "🔬 Analyze Sequence", variant="primary", size="lg", elem_classes=["analyze-btn"]
+                    )
+
+                gr.Examples(
+                    examples=examples, inputs=[sequence_input], label="📚 Example Sequences", elem_id="examples"
+                )
+
+            with gr.Column(scale=1, elem_classes="result-column"):
+                # Results Section
+
+                gr.Markdown("## 📊 Analysis Results")
+
+                result_output = gr.Markdown(
+                    value="✨ Enter a sequence and click 'Analyze Sequence' to see detailed results and visualizations.",
+                    elem_id="results",
+                )
+
+                # Enhanced plot component
+                plot_output = gr.Plot(label="📈 Probability Distribution", value=None, elem_id="probability-plot")
+
+        # Footer Section
+        with gr.Row():
+            gr.Markdown(
+                """
+                ## 🚀 About ChimeraLM
+
+                **Advanced Features:**
+                - ⚡ **High Performance**: Optimized for speed and accuracy
+                - 🎯 **Binary Classification**: Distinguishes biological vs chimeric sequences
+                - 📏 **Long Sequences**: Handles up to 32,768 nucleotides
+                - 🤖 **Pre-trained Model**: Ready-to-use with `yangliz5/chimeralm`
+
+                **Technical Specifications:**
+                - **Model Type**: Binary Sequence Classifier
+                - **Input**: DNA sequences with standard nucleotides
+                - **Output**: Classification + confidence scores
+                - **Training**: Whole genome amplification artifact detection
+
+                ---
+
+                **📖 Citation:**
+                ```
+                @software{chimeralm2025,
+                  title={ChimeraLM: A genomic language model to identify chimera artifacts},
+                  author={Li, Yangyang, Guo, Qingxiang and Yang, Rendong},
+                  year={2025},
+                  url={https://github.com/ylab-hi/ChimeraLM}
+                }
+                ```
+
+                **🔗 Links:**
+                - [GitHub Repository](https://github.com/ylab-hi/ChimeraLM)
+                - [Model Hub](https://huggingface.co/yangliz5/chimeralm)
+                - [Documentation](https://github.com/ylab-hi/ChimeraLM#readme)
+                """,
+                elem_classes="footer-section",
+            )
+
+        # Connect the button click
+        predict_btn.click(fn=predict_sequence, inputs=[sequence_input], outputs=[result_output, plot_output])
+
+    return interface
+
+
+if __name__ == "__main__":
+    logger.info("🚀 Starting ChimeraLM Web Interface...")
+    interface = create_interface()
+    interface.launch(share=False)

requirements.txt

@@ -0,0 +1,33 @@
+# Core dependencies for ChimeraLM web UI
+chimeralm>=1.0.4
+
+# Deep Learning
+torch==2.5.1
+torchvision>=0.20.1
+torchaudio>=2.5.0
+lightning>=2.4.0
+torchmetrics>=1.6.0
+
+# ML/NLP
+transformers>=4.47.1
+datasets>=3.2.0
+einops>=0.8.0
+evaluate>=0.4.3
+
+# Bioinformatics
+pysam>=0.22.1
+pyfastx>=2.2.0
+
+# Web UI
+gradio>=5.0.0
+plotly>=5.24.0
+
+# Configuration
+hydra-core>=1.3.2
+omegaconf>=2.3.0
+
+# Utilities
+rich>=13.9.4
+typer>=0.15.1
+joblib>=1.5.2
+hf-xet>=1.1.10