Chemical Coverage Optimization

Reconstruct the targeted peak extraction strategy for rescuing missed features

Summary

This skill applies a targeted peak extraction strategy to LC-MS feature tables in order to rescue chemical features that traditional peak extraction algorithms fail to detect, thereby expanding the chemical substance annotation coverage without reliance on additional chemical standards.

When to use

Your LC-MS feature table is incomplete or has low chemical coverage because traditional peak extraction algorithms (e.g., standard XCMS workflows) systematically miss features at m/z and retention time positions corresponding to known suspect compounds in your database. Use this skill when you have a suspect compound database (e.g., T3DB) and want to augment an existing feature table by deliberately extracting peaks at database-specified m/z–RT coordinates that were not recovered in the initial peak picking pass.

When NOT to use

• Your input is already a high-coverage annotated feature table from a tool specifically designed for suspect screening (e.g., already processed through EISA); targeted extraction adds noise rather than coverage.
• You lack a validated suspect compound database or cannot justify the m/z and RT coordinates for your compounds of interest.
• Your LC-MS data has poor peak quality, high chemical noise, or insufficient mass accuracy (>10 ppm) to reliably distinguish targeted peaks from background.

Inputs

• LC-MS raw data or preprocessed feature table (mzML, netCDF, or vendor format)
• Suspect compound database in .xlsx or .csv format with columns: NAME, PrecursorMZ, ProductMZ, Intensity, RT (optional), ID
• Existing feature table from traditional peak extraction (e.g., XCMS output)

Outputs

• Augmented feature table with targeted-extraction-recovered features flagged and integrated
• Metadata table linking rescued features to suspect compound identities and database entries
• Visualization and filtering results from Shiny interface (optional)

How to apply

Load your LC-MS raw data (or preprocessed feature table) and a structured suspect compound database (columns: NAME, PrecursorMZ, ProductMZ, Intensity, RT, ID) into the EISA-EXPOSOME R Shiny environment. For each suspect compound in the database, use its known m/z and retention time as targeted extraction coordinates to search the raw LC-MS data for peaks that traditional algorithms missed. Integrate successfully recovered peaks back into your original feature table, labeling them as targeted-extraction rescues. Apply optional filtering and visualization through the Shiny interface to validate extracted features and remove false positives based on peak intensity and spectral coherence.

Related tools

• R Shiny (EISA-EXPOSOME interface) (Interactive platform for applying targeted peak extraction, visualizing recovered features, and filtering results by intensity and retention time) — https://github.com/Lab-XUE/EISA-EXPOSOME
• T3DB (compiled database file) (Pre-formatted suspect compound database (.xlsx) providing m/z, RT, and product ion information for targeted extraction queries) — https://github.com/Lab-XUE/EISA-EXPOSOME

Evaluation signals

• Number of features rescued by targeted extraction exceeds zero and is consistent with the number of suspect compounds in the database with valid m/z–RT coordinates.
• Rescued features appear in the augmented feature table with correct m/z ± mass accuracy tolerance (typically <5 ppm for high-resolution MS) and RT within expected retention window (±0.5 min).
• Intensity values of rescued peaks are above instrument noise floor and display reasonable product ion fragmentation patterns consistent with database records.
• Downstream chemical annotation coverage increases (e.g., more compounds identified in your sample) after targeted extraction compared to traditional peak extraction alone.
• Visualization in Shiny interface confirms that rescued features do not represent systematic artifacts or repeated extractions of the same underlying peak.

Limitations

• The skill relies on the accuracy and completeness of the suspect compound database; missing or incorrect m/z or RT values will prevent recovery of true features or lead to false extractions.
• Targeted extraction may recover spurious peaks if the LC-MS data has high background noise, co-eluting contaminants, or poor mass calibration; manual validation or stricter intensity thresholds may be necessary.
• RT values in the database may shift between instruments or chromatographic methods; the workflow assumes RT data are either provided or can be reliably predicted or empirically validated.
• The strategy is designed for targeted screening of known suspect compounds; it does not discover novel or unanticipated chemical features not in the database.

Evidence

• [intro] Targeted peak extraction strategy rescues features that cannot be extracted by traditional peak extraction algorithms: "used a targeted peak extraction strategy to rescue features that cannot be extracted by traditional peak extraction algorithms, improving the coverage of chemica substance annotation"
• [other] Workflow input and output specifications: "Load LC-MS feature/peak data and the suspect database (e.g., T3DB in .xlsx format) into memory. Apply the targeted peak extraction strategy to identify and extract peaks at m/z and retention time"
• [readme] Database format requirements: "your file (.xlsx /.csv) must contain the following columns:|NAME|PrecursorMZ|ProductMZ|Intensity|RT|ID|, RT is not essential."
• [readme] Platform and tool delivery: "We provide a Rshiny program for EISA-EXPOSOME, which runs with the interface shown below, and you can filter the results according to the visualisation interface"
• [intro] Benefit and coverage improvement: "can help reduce the dependence on chemical standards in traditional chemical analysis and significantly enhance the chemical coverage"