Library search
Library search names a discovered spectrum. It runs from Compound Discovery after deconvolution or after a SIM/MRM component is built. The libraries are saved on the method.
Libraries dialog
Libraries in Compound Discovery opens the list. Each row is one file:
| Column | Meaning |
|---|---|
| Path | ChemStation .L folder, or MassHunter .mslibrary.xml. Anything else is rejected. |
| Use | Include this library in the search. |
| Priority | When two hits have the same rank, the higher priority library wins. |
| Prescreen | Slow, Standard, or Fast for that library. |
You can sort the columns. Slow does not drop entries before the full score. Standard and Fast drop entries that cannot explain the spectrum. Faster prescreen keeps fewer candidates and can miss a difficult match. The choice is stored with the library on the method.
There is no .msp row. The hit list still shows a match probability for each candidate. The file you select is .L or .mslibrary.xml.
Spectrum origin
Each search knows how the query spectrum was made:
| Origin | When | Prescreen (Standard) | How hits are ranked |
|---|---|---|---|
| Raw apex | Spectrum at the peak top, extras included | Looser (more unexplained peaks allowed) | 35% Match + 65% RMatch |
| Background-subtracted | Apex minus nearby background | Stricter | Match only |
| Deconvoluted | Component spectrum from SCAN deconvolution | Stricter | Match only |
Fast is stricter than Standard for the fraction of the spectrum that must be explained. Slow turns the prescreen off.
A cleaned spectrum (background-subtracted or deconvoluted) is ranked by Match, because extra noise peaks should already be gone. A raw apex is ranked mostly by RMatch, because extra sample peaks are expected.
Scores
| Column | Range | What it emphasizes |
|---|---|---|
| Match | 0–999 | Forward match. The whole sample spectrum against the library. Extra sample peaks and missing library peaks both lower it. Best for a clean, deconvoluted spectrum. |
| RMatch | 0–999 | Reverse match. Only library peaks count, so extra sample peaks are ignored. A missing library peak still hurts. If RMatch is much higher than Match, the spectrum is probably impure. |
| Prob | relative | How unique this hit is among the hits for this spectrum. It stays low when several library compounds look alike, even if Match is high. |
| PBM % | percent | McLafferty probability-based match. Rare, high-mass peaks count more than common low-mass peaks. A missing minor peak is tolerated. Useful on a noisy or partial spectrum. |
The compound grid shows the top hit: Lib Name, CAS, Match, RMatch, Prob, PBM %.
Reloading and errors
Libraries are loaded once and reused. If the file on disk changes, the next search reloads it. A missing file or a file that is not a .L or .mslibrary.xml is skipped. The logbook records the error (Tools → Logbook). The search continues with the libraries that did load.
See SCAN deconvolution and Compound Discovery.