In OpenLAB CDS v2.x, the threshold is calculated automatically for every data point. The sharpness of the analysis is adjusted using the sensitivity percentage.
The sensitivity is therefore not a fixed threshold value. Modifying the sensitivity does not change the threshold value with a linear relationship, as the threshold values are not identical at every raw data point. As an example, on the figure below, if the threshold value at 2.950 min is 998, it does not necessarily mean that the threshold would be the same at 2.960 min. Every data point have its own threshold calculated, leading to a threshold curve across the peak.
The peak purity sensitivity is then applied to the threshold curve. In Data Analysis, a logarithmic transformation of the threshold and similarity curves is drawn below the peaks. The threshold curve is then a flat line between the pure and impure regions. The similarity curve (shown in Peak Details) is drawn using the formula as described under Impurity evaluation. This similarity curve is then showing a distribution of positive values (pure data points), and negative values (impure data points) across the peak, where the 0 represents the threshold limit.
For example:
Data point with a match factor of 990 and a threshold of 980:
Ratio = log( (1000 - 980) / (1000 - 990) ) = log(2) = +0.3
This raw data point meets the purity criteria.
Data point with a match factor of 970 and a threshold of 990:
Ratio = log( (1000 - 990) / (1000 - 970) ) = log(0.33) = -0.48
This raw data point does not meet the purity criteria.
Decreasing or increasing the sensitivity means that the profile of the calculated threshold curve is changing. The full range for sensitivity is from 0 to 100 %, where the default calculated threshold is at 50 %.
Because the display of the similarity curve is nonlinear (but logarithmic), the threshold will not map to a one to one relationship from one data point to another. If the sensitivity is increased/decreased by 20 %, the threshold curve is moved up or down, and its amplitudes change as well. The raw data point thresholds are not changed by +/- 20 %.
Sensitivity | Threshold |
|---|---|
|
0 % |
Lowest possible value = 0 |
|
0 <= s <= 100 |
Calculated threshold curves (with a reference curve at 50%) |
|
100 % |
Highest possible value = 1000 |
Low sensitivity - this peak is considered pure |
Default sensitivity - this peak is considered impure |
High sensitivity - this peak is considered impure |
A peak is flagged as impure when one data point is under its threshold, even if the overall UV peak purity factor is close to 1000.
base-id: 11016935691
id: 11016935691