The following table describes the parameters for the ReSpect deconvolution algorithm on the Component Detection page.
See Best results with the ReSpect algorithm for information on selecting ideal ReSpect deconvolution parameter values.
Parameter | Description |
---|---|
Deconvolution Algorithm | Select the ReSpect (Isotopically Unresolved) option. |
Parameters (ReSpect) | |
Deconvolution Results Filter | |
Output Mass Range | Specifies the range for the displayed masses on the x axis of the plots in the Deconvoluted Spectra panes on the Process and Review page. |
Deconvoluted Spectra Display Mode | Specifies the mode to display the Deconvoluted Spectra panes on the Process and Review page:
|
Charge State Distribution | |
Deconvolution Mass Tolerance | Specifies the global allowable error for the m/z values of peaks in a charge state series as they appear in the input spectrum. Also compensates for calibration errors and the effects of local noise, peak overlaps, and other sources of mismatches between the model and the actual peak profiles. This parameter is important in the ReSpect algorithm, because it controls the tolerance for peaks, in m/z, when the algorithm uses these peaks to fit to a particular component. As you increase the value of this parameter, the ReSpect algorithm expands the plausible set of charge state peaks, but at the same time it correspondingly increases the false positive rate. The ReSpect algorithm can exclude these false positives to some degree. For most purposes, the ppm setting provides better results than the Da setting. A ppm setting between 20 and 50 is a good starting point. Select the unit for this parameter:
|
Choice of Peak Model | |
Choice of Peak Model | Specifies the appropriate peak model for the data. The expected peak shapes for nucleotides are different from those for proteins and peptides because of their different isotopic composition.
|
Resolution at 400 m/z | Defines the resolution of the source spectrum at an m/z value of 400. Select one of these options:
|
Generate XIC for Each Component | |
Calculate XIC (checkbox) | When selected, the application calculates the extracted ion chromatogram (XIC) for each detected component. Enabling this parameter can result in a much longer analysis time, so you might avoid using it with complex data or with data where XIC data is unnecessary. For quantitative analysis, use the abundance trace for a given component for more accurate and reliable quantification calculations. Abundance traces are generated for experiments using either the Xtract of ReSpect algorithms with the sliding window deconvolution method. |
Advanced Parameters (ReSpect) Visible only when you select the Show Advanced Parameters checkbox; these parameters typically do not need to be modified. | |
Charge State Distribution | |
Model Mass Range | Specifies the required mass range for the model from the minimum (lowest) to the maximum (highest) allowable mass. The ReSpect algorithm rejects potential components with masses less than the minimum or greater than the maximum mass specified.. |
Charge State Range | Specifies the required charge state range for the model from the minimum (lowest) to the maximum (highest) allowable charge state. The ReSpect algorithm rejects potential components with fewer than the minimum or greater than the maximum number of charge states. |
Minimum Adjacent Charges (low & high model mass) | Specifies the minimum number of charge state peaks that must appear in a row for components of the low model mass and of the high model mass. For example, if the specified values for Minimum Adjacent Charges are 6 and 10 and the model mass range is 10,000 to 160,000, a component with a mass of 10,000 would require at least 6 adjacent charge states, and a component with a mass of 160,000 would require at least 10 adjacent charges states. The performance of the ReSpect algorithm improves as the number of adjacent charge states increases. Research suggests that this algorithm always yields reliable results if the minimum numbers of allowed adjacent charge states at the low and high end of the m/z range are 6 and 10, respectively. Therefore, these numbers are the default parameters. |
Noise Parameters | |
Rel. Abundance Threshold (%) | Specifies a relative abundance threshold below which the application filters out data for analysis. This option sets a threshold as a percentage of the most abundant component in the spectrum. The most abundant peak in the deconvoluted spectrum has a relative abundance of 100 percent, and all other peaks are calculated relative to that one. In the Results table on the Process and Review page, the application shows only those components that are greater than or equal to this relative abundance threshold in the deconvoluted spectrum. For example, if the highest peak has an absolute abundance of 1000 and the relative abundance threshold is 1 percent, the application filters out all peaks below an absolute abundance of 10. For this value, 0% displays all results and 100% displays only the most abundant component. |
Deconvolution Quality | |
Quality Score Threshold | Specifies a minimum quality score that a component must have to be displayed in the Results table. The algorithm discards components that have a score beneath this threshold. |
Peak Filter Parameters | |
Peak Detection Minimum Significance Measure | Specifies a significance level, in standard deviations, that determines whether the ReSpect algorithm discards a peak as a noise feature (spectrum noise from the background) or retains it as a legitimate peak. The ReSpect algorithm retains peaks equal to or greater than this selected significance level. The higher the significance level, the stricter the filter. |
Peak Detection Quality Measure | Removes noise and irrelevant features from the list of peaks. Noise is calculated from the spectrum background. This parameter is important in the ReSpect algorithm because it controls how plausible a peak must be before the algorithm uses it in a deconvolution. Reducing the confidence level for this parameter increases the number of peaks but allows more noise. Tests suggest that a confidence level of 95 percent works well in most cases and the value can be relaxed down to 68 percent if necessary. A confidence level of 99 percent is rarely required. A confidence level of 50 percent or less might produce an unacceptable number of false positives.
|
Choice of Peak Model | |
Target Mass | Specifies an expected target mass, in Da, to use in calculating the peak model. This parameter is critical but does not have to be exact; a value within 5 and 10 percent of the actual target is sufficient for best performance. For samples where the range of masses is broad, choose a mass somewhere in the middle of the range. For example, if your sample contains immunoglobulin G with two light chains (20 kDa), two heavy chains (50 kDa), and the intact antibody in the same sample, choose 75 kDa as the target mass. When you modify the maximum value of the Output Mass Range, the application automatically updates the Target Mass value to match the modified value. However, modifying the Target Mass value does not affect the Output Mass Range values. Previous versions of the BioPharma Finder and Protein Deconvolution applications used a default value of 150,000 for this parameter. |
Peak Model Parameters | |
Number of Peak Models | Controls the resolution of the peak modeling process by dividing the observed m/z range into a uniformly spaced set of regions equal to this number. The application generates a single peak model for each of these regions on the basis of the observed m/z value and instrument resolution at the midpoint of each region. |
Left/Right Peak Shape | Defines the sharpness of a peak. |
Specialized Parameters | |
Peak Model Width Factor | Specifies the adjustment to the peak width in the model used by the algorithm. A larger value makes the peak width wider and vice versa. |
Intensity Threshold Scale | Specifies how intense a possible charge state peak must be to be included in the wings of a charge state distribution.
Reducing this threshold value can increase sensitivity at the expense of a possible increase in the false-positive rate for weak-intensity low-scoring components. Previous versions of the BioPharma Finder and Protein Deconvolution applications used a default value of 0.01 for this parameter. |
Deconvolution Parameters | |
Noise Compensation | When selected (default), the ReSpect algorithm improves signal detection when the noise level varies across the data. |
Charge Carrier | Specifies the adduct ions introduced during the electrospray ionization (ESI) process. Adduct ions convert neutral molecules into ions.
In negative mode, these ions correspond to losses rather than adducts. |
Negative Charge | When cleared, indicates the data was acquired in positive charge mode during the electrospray ionization (ESI) process. When selected, indicates the data was acquired in negative charge mode. MS spectra of samples containing nucleotides (such as DNA or RNA) are often collected in negative charge mode. Do not select the Negative Charge checkbox if your data was acquired in positive mode. Results will not be usable. |