LCMS Adduct Finder

Automated Targeted Feature Extraction & Adduct Verification Tool for LC-MS Data.

This Python tool is designed for the targeted analysis of LC-MS data. By providing a list of Chemical Formulas, it automatically performs a comprehensive scan for various adduct forms (e.g., [M+H]+, [M+Na]+). It extracts Extracted Ion Chromatograms (EIC) and rigorously evaluates peak quality using Gaussian fitting to determine the reliability of the detected signals.

---

Key Features

• Targeted Extraction: Instantly converts chemical formulas (e.g., C6H12O6) into target m/z values, enabling precise extraction of specific metabolites or compounds.
• Multi-Adduct Verification:
  • Automatically scans for 14+ different adduct types (Monomers, Dimers, Na/NH4 adducts, etc.) simultaneously.
  • Helps confirm the identity of a substance by checking if multiple adducts elute at the same Retention Time (RT).
• Peak Quality & Existence Check:
  • Gaussian Scoring: Fits a Gaussian curve to the raw peak data and calculates an R² score.
  • Distinguishes high-quality peaks ("Excellent/Good") from noise or irregular shapes ("Poor/Noise").
• Precision Mass Calculation: Uses high-precision logic considering electron mass: $$m/z = \frac{(M \times n + \Delta) - (Charge \times m_e)}{|Charge|}$$
• Visual Inspection (Optional): Saves EIC plots as PNG images with Gaussian fit overlay.
• MS2 Matching: Links MS1 features to MS2 events for additional confirmation.

---

Prerequisites

This tool uses pythonnet and fisher-py to read Thermo .raw files, which requires a .NET runtime. Installation varies by operating system:

Windows

.NET Framework is typically pre-installed on Windows 10/11. If needed, install the .NET Runtime (version 4.7.2 or later recommended).

Linux (Ubuntu/Debian)

Install Mono runtime:

sudo apt update
sudo apt install -y mono-complete

For other distributions, see the Mono installation guide.

macOS

Install Mono using Homebrew:

brew install mono

Or download the installer from the Mono project website.

---

Installation

From PyPI (recommended)

pip install lib_eic

With YAML configuration support:

pip install lib_eic[yaml]

From Source

git clone https://github.com/SNUFML/lib_eic.git
cd lib_eic
pip install -e .

Dependencies

• pandas, openpyxl (Excel I/O)
• molmass (mass calculations)
• scipy, numpy (numerical processing)
• matplotlib (plotting)
• fisher-py, pythonnet (Thermo .raw file reading)

---

Supported Adducts

The tool automatically detects the ionization mode (Positive/Negative) and scans for the following adducts:

Mode	Adduct Types
Positive (+)	[M+H]+, [M+Na]+, [M+NH4]+, [M+ACN+H]+, [2M+H]+, [M-H2O+H]+, etc.
Negative (-)	[M-H]-, [M+FA-H]-, [M-H2O-H]-, [2M-H]-, etc.

---

Usage

Quick Start

1. Place your Thermo .raw files in ./raw folder
   • Nested layouts are supported, e.g. ./raw/{RP,HILIC}/{1st,2nd}/*.raw
2. Create an input Excel file (file_list.xlsx) with your compound list
3. Run the tool:

lib_eic
# or
python -m lib_eic

Input Excel Format

lib_eic supports two input formats (auto-detected by column names).

A) Direct m/z format (recommended for EIC plot generation)

The Excel file contains separate sheets for chromatography modes:

• RP (Reverse Phase)
• HILIC (Hydrophilic Interaction Liquid Chromatography)

The Excel file may contain merged cells in row 1; headers/data start from row 2.

num	File name	mixture	Compound name	Polarity	m/z
1	Library_POS_Mix121	121	Spermine	POS	203.223
2	Library_POS_Mix121	121	Putrescine	POS	89.107
3	Library_NEG_Mix121	121	Glucose	NEG	179.056

• num: Optional ordering number (used to prefix plot filenames for easier sorting)
• File name: Partial raw filename prefix used for matching (e.g., matches File name.raw, File name_2nd.raw, ...)
• mixture: Mixture identifier (used in plot filenames)
• Compound name: Display name for plots and Excel output
• Polarity: POS or NEG
• m/z: Direct target m/z value

EIC plots are saved under: EIC_Plots_Export/{LC mode}/{Polarity}/{File name}/[{num}_]{Compound name}_{Polarity}_{mixture}{suffix}.png

Notes:

• For direct m/z input (separate RP/HILIC sheets), if --raw-folder contains an {LC mode} subfolder, the tool searches that first.
• If raw files are further split by run folders (e.g. 1st/, 2nd/), the run label is carried into the output (and plot filenames) to avoid overwrites.

B) Formula-based format (legacy)

RawFile	Mode	Formula
sample_01.raw	POS	C6H12O6
sample_01.raw	POS	C10H16N5O13P3
sample_02.raw	NEG	C6H12O6

• RawFile: Filename (extension can be omitted; .raw is appended if missing)
• Mode: POS or NEG (uppercase)
• Formula: Chemical formula to analyze

---

Configuration

Option 1: Command Line Arguments

lib_eic --raw-folder ./my_raw_files --input compounds.xlsx --output results.xlsx --ppm 10.0 -v

Common options:

• --raw-folder: Path to folder containing .raw files (default: ./raw)
• --input: Input Excel file path (default: file_list.xlsx)
• --output: Output Excel file path (default: Final_Result_With_Plots.xlsx)
• --pivots: Enable per-target pivot table sheets
• --no-pivots: Disable per-target pivot table sheets (default; faster for large runs)
• --ppm: Mass tolerance in ppm (default: 10.0)
• --no-plots: Disable EIC plot generation
• --no-fitting: Disable Gaussian fitting
• --no-ms2: Disable MS2 indexing/matching
• --workers N: Number of worker processes (default: auto; use 1 for sequential)
• --sequential: Force sequential processing (equivalent to --workers 1)
• --no-progress: Disable the tqdm progress bar
• -v, --verbose: Enable verbose output
• --help: Show all available options

Performance notes:

• Parallelism is file-level (one worker per raw file); if you only have 1 raw file, speedup is limited.
• If CPU usage stays low, the run is likely bottlenecked by disk I/O (.raw reads) or output writing (Excel/plots); try fewer workers and/or an SSD, and consider --no-plots and leaving pivot sheets disabled.

Option 2: YAML Configuration File

Generate a default configuration template:

lib_eic --generate-config config.yaml

Then edit and use:

lib_eic --config config.yaml

Example config.yaml:

raw_data_folder: "./raw"
input_excel: "file_list.xlsx"
input_sheets: ["RP", "HILIC"]
output_excel: "Final_Result_With_Plots.xlsx"
include_pivot_tables: false
show_progress: true
num_workers: 0          # 0 = auto, 1 = sequential, N = N workers
parallel_mode: "auto"   # "auto", "sequential", "file" (file-level multiprocessing)
ppm_tolerance: 10.0
min_peak_intensity: 100000
enable_fitting: true
enable_plotting: true
enable_ms2: true
export_plot_folder: "EIC_Plots_Export"
area_method: "sum"  # or "trapz"
ms2_match_mode: "rt_linked"  # or "global"

Option 3: Python API

from lib_eic.config import Config
from lib_eic.processor import process_all

# Create configuration
config = Config(
    raw_data_folder="./raw",
    input_excel="file_list.xlsx",
    output_excel="results.xlsx",
    ppm_tolerance=10.0,
    enable_plotting=True,
    enable_fitting=True
)

# Run analysis
process_all(config)

Advanced: Direct Module Access

from lib_eic.chemistry.mass import get_exact_mass, calculate_target_mz
from lib_eic.chemistry.adducts import get_enabled_adducts
from lib_eic.analysis.eic import build_targets, extract_eic
from lib_eic.analysis.fitting import fit_gaussian_and_score
from lib_eic.io.raw_file import RawFileReader

# Calculate exact mass
mass = get_exact_mass("C6H12O6")  # ~180.0634

# Get enabled adducts for positive mode
adducts = get_enabled_adducts("POS")

# Build targets from formulas
targets = build_targets(["C6H12O6"], adducts)

# Read raw file and extract EIC
with RawFileReader("sample.raw") as reader:
    rt, intensity = reader.get_chromatogram(target_mz=181.0707, ppm=10.0)

---

Output

1. Excel Report (Final_Result_With_Plots.xlsx)

• All_Features Sheet: Complete per-target results table (includes targets that were not reported as features)

  • Adds EICGenerated (whether chromatogram extraction returned data) and FilteredOut (below --min-intensity)
  • Formula-based: RawFile, Mode, Formula, Adduct, mz_theoretical, RT_min, Intensity, Area, GaussianScore, PeakQuality, HasMS2, EICGenerated, FilteredOut
  • Direct m/z: num, RawFile, File name, lc_mode, mixture, Compound name, Polarity, mz_target, RT_min, Intensity, Area, GaussianScore, PeakQuality, HasMS2, EICGenerated, FilteredOut

• Per-Target Sheets: Pivot tables for each Formula / Compound name

  • Area Table: Peak areas across samples and adducts
  • Retention Time Table: RT consistency verification across adducts

2. EIC Plots (EIC_Plots_Export/)

Visual validation of detected peaks (when plotting is enabled):

• Direct m/z: EIC_Plots_Export/{Polarity}/{File name}/{Compound name}_{Polarity}_{mixture}{suffix}.png
• Blue Line: Raw EIC data
• Red Marker: Apex RT indicator
• Dashed Line: Gaussian fit curve (when fitting is enabled)

---

Quality Scoring System

The Gaussian fitting provides an R² score to assess peak quality:

R² Score	Label	Interpretation
> 0.8	Excellent	High-quality Gaussian peak, reliable signal
0.5-0.8	Good	Adequate fit, signal is likely real
< 0.5	Poor	Irregular shape, may be artifact or noise
N/A	Not Fitted	Fitting was disabled

---

Testing

Run the test suite:

pytest tests/
pytest -v tests/  # verbose output
pytest --cov=lib_eic tests/  # with coverage report

---

Project Structure

lib_eic/
├── chemistry/          # Mass calculations and adduct definitions
│   ├── mass.py         # Exact mass and m/z calculations
│   └── adducts.py      # Adduct type definitions
├── analysis/           # Core analysis modules
│   ├── eic.py          # EIC extraction and target building
│   ├── fitting.py      # Gaussian peak fitting
│   └── ms2.py          # MS2 precursor matching
├── io/                 # Input/output operations
│   ├── raw_file.py     # Thermo .raw file reader
│   ├── excel.py        # Excel I/O
│   └── plotting.py     # EIC visualization
├── config.py           # Configuration management
├── processor.py        # Main processing pipeline
├── cli.py              # Command-line interface
└── validation.py       # Input validation

---

Limitations

• Supported File Format: Thermo .raw files only (via fisher-py)
• mzML Not Supported: Use vendor-specific raw files for best results
• Platform: Requires .NET runtime (Windows native, or Mono on Linux/macOS)

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Author

Jihyun Chun (jihyun5311@snu.ac.kr)

Repository: https://github.com/SNUFML/lib_eic