# PASC-Project

**Repository Path**: haolw1976/PASC-Project

## Basic Information

- **Project Name**: PASC-Project
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-01-28
- **Last Updated**: 2026-01-28

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# PASC Analysis: Multi-Omics Approaches to Long COVID Research

This repository contains R scripts for analyzing Post-Acute Sequelae of COVID-19 (PASC) using multiple analytical approaches including immunological assays, transcriptomics, and machine learning methods.

## Table of Contents
- [Overview](#overview)
- [Requirements](#requirements)
- [Installation](#installation)
- [Data Structure](#data-structure)
- [Analysis Scripts](#analysis-scripts)
- [Usage](#usage)
- [Output Files](#output-files)
- [Citation](#citation)
- [Contact](#contact)

## Overview

This repository provides a comprehensive analytical pipeline for PASC research, implementing:

- **Immunological Analysis**: ELISPOT and neutralizing antibody assay analysis with mixed-effects modeling
- **Transcriptomic Analysis**: RNA-seq analysis using batch correction, differential expression, and pathway enrichment
- **Machine Learning**: Random Forest classification for PASC prediction
- **Network Analysis**: Weighted Gene Co-expression Network Analysis (WGCNA) for module identification
- **Pathway Analysis**: Single-sample Gene Set Enrichment Analysis (ssGSEA)

## Requirements

### R Version
- R >= 4.0.0

### Required R Packages

```r
# Core data manipulation and visualization
install.packages(c("tidyverse", "ggplot2", "dplyr", "magrittr"))

# Mixed-effects modeling
install.packages(c("lme4", "lmerTest", "broom.mixed"))

# Bioconductor packages
if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install(c(
    "Biobase", "limma", "DESeq2", "genefilter", 
    "GSEABase", "GSVA"
))

# Machine learning
install.packages(c("randomForest", "caret", "pROC"))

# Network analysis
install.packages("WGCNA")

# Additional utilities
install.packages(c("readxl", "phateR"))
```

## Installation

1. Clone this repository:
```bash
git clone https://github.com/yourusername/pasc-analysis.git
cd pasc-analysis
```

2. Install required packages (see Requirements section above)

3. Prepare your data files according to the Data Structure section below

## Data Structure

### Expected Input Files

#### For Immunological Analysis:
- `MassCPRElispot.csv`: ELISPOT assay data with columns:
  - First column: Participant ID
  - `STATUS` column: "WEAK" or "STRONG"
  - Assay columns: `Elispot.WH`, `Elispot.Delta`, `Elispot.B.1.1.529`

- `MassCPRNab_Elispot.csv`: Neutralizing antibody data with columns:
  - First column: Participant ID
  - `STATUS` column: "WEAK" or "STRONG"  
  - Assay columns: `nAb.WH`, `nAb.DL`, `nAb.BA.1`

#### For Transcriptomic Analysis:
- `PASC.RData`: ExpressionSet object for cohort 1
- `PASC2025.Rdata`: ExpressionSet object for cohort 2
- `Cohort1.csv` and `Cohort2.csv`: Metadata files with sample information
- `features.csv`: Gene annotation file

#### For Pathway Analysis:
- `GSEA.F.vs.M.lc.3-6M.csv`: GSEA results file with pathway information
- `IL6_JAK_STAT.csv`: Gene lists for specific pathways of interest

### Data Format Requirements

**ExpressionSet Objects** should contain:
- Expression matrix (genes × samples)
- Phenotype data with columns: `SampleID`, `SubjectID`, `Time`, `PASCSTATUS`
- Feature data with gene annotations

**CSV Files** should be properly formatted with headers and consistent naming conventions.

## Analysis Scripts

### 1. `MixedEffectModel.PASC.Elispot.R`
**Purpose**: Analyze ELISPOT assay data using mixed-effects models

**Key Features**:
- Log10 transformation of SFU values
- Mixed-effects modeling accounting for participant-level random effects
- Statistical comparison between WEAK vs STRONG groups
- Automated plotting with p-values

**Usage**:
```r
source("MixedEffectModel.PASC.Elispot.R")
```

### 2. `MixedEffectModel.PASC.nAbs.R`
**Purpose**: Analyze neutralizing antibody titers using mixed-effects models

**Key Features**:
- Handles limit of detection (LOD) values
- Log10 transformation of titers
- Mixed-effects statistical modeling
- Faceted visualization across variants

**Usage**:
```r
source("MixedEffectModel.PASC.nAbs.R")
```

### 3. `RandomForest_PASC_Combined_Cohorts.R`
**Purpose**: Machine learning classification of PASC status using RNA-seq data

**Key Features**:
- Batch correction using `limma::removeBatchEffect()`
- Random Forest classification
- Feature importance analysis
- ROC curve generation and performance metrics
- Pathway-specific analysis

**Usage**:
```r
source("RandomForest_PASC_Combined_Cohorts.R")
```

### 4. `ssgsea_pasc.R`
**Purpose**: Single-sample Gene Set Enrichment Analysis

**Key Features**:
- Pathway enrichment scoring for individual samples
- Integration with GSEA results
- Standardized sample naming
- Expression matrix preprocessing

**Usage**:
```r
source("ssgsea_pasc.R")
```

### 5. `WGCNA_pasc.R`
**Purpose**: Weighted Gene Co-expression Network Analysis

**Key Features**:
- Network construction and module detection
- Module-trait relationship analysis
- Soft thresholding power selection
- Temporal analysis (Acute, Mid, High time points)
- Focus on specific pathway genes (e.g., JAK/IL6 signaling)

**Usage**:
```r
source("WGCNA_pasc.R")
```

## Usage

### Basic Workflow

1. **Prepare Data**: Ensure all required input files are in the working directory
2. **Set Parameters**: Modify file paths and parameters in scripts as needed
3. **Run Analysis**: Execute scripts in order based on your research questions

### Example Analysis Pipeline

```r
# 1. Immunological analysis
source("MixedEffectModel.PASC.Elispot.R")
source("MixedEffectModel.PASC.nAbs.R")

# 2. Transcriptomic analysis
source("RandomForest_PASC_Combined_Cohorts.R")

# 3. Pathway analysis
source("ssgsea_pasc.R")

# 4. Network analysis
source("WGCNA_pasc.R")
```

### Customization

**Key Parameters to Modify**:
- File paths in each script
- Assay column names to match your data
- Time point definitions
- Pathway gene lists
- Random Forest parameters (ntree, etc.)
- WGCNA soft threshold power

## Output Files

### Immunological Analysis
- `elispot_scatter_all_assays.pdf`: Combined ELISPOT results
- `neutralization_scatter_all_assays.pdf`: Combined neutralization results
- Individual assay PDFs for each variant

### Machine Learning Analysis
- `*_FeatureImportance.pdf`: Top predictive genes
- `*_ROC.pdf`: ROC curves with AUC scores  
- `*_Top_Predictive_allgenes.csv`: Gene importance rankings
- `combined.expr.matrix.csv`: Batch-corrected expression data

### Pathway Analysis
- `FvsM_lc_3-6M_ssgseaMatrix_C2.csv`: ssGSEA enrichment scores

### Network Analysis
- `Tree_jakMid_filtered.pdf`: Gene dendrogram and module colors
- `JakMid_filteredHEAT.pdf`: Module-trait relationships heatmap
- `JakMid_filteredPROF.pdf`: Expression profiles by module
- `gene_modules.txt`: Gene-to-module assignments

## Statistical Methods

- **Mixed-Effects Models**: Account for repeated measures and participant-level variation
- **Batch Correction**: Remove technical variation between cohorts using `limma`
- **Random Forest**: Non-linear classification with feature importance
- **WGCNA**: Identify co-expressed gene modules and their relationships to phenotypes
- **ssGSEA**: Calculate pathway enrichment scores for individual samples

## Citation

If you use this code in your research, please cite:

```
[TBD]
```

## Contact

For questions or issues, please contact:
- **Author**: [Malika Aid]
- **Email**: [maid@bidmc.harvard.edu]
- **Institution**: [BCGE LAB]

## License

**Note**: This repository is actively maintained. Please check for updates and report any issues through GitHub Issues.