(Summary of the calibration sample details from the NOCIt software technical manual)
Calibration data requires diversity of genotypes and masses to provide information for modeling, including data on peak heights, peak height ratios, degradation, and stutter. The number of unique samples typically ranges from 20 – 30, but can be more or fewer samples, depending on diversity and heterozygosity. In general, the more unique samples the better as some samples may have variation due to binding site issues or amplification efficiency. The electropherograms should be analyzed at the lowest RFU (usually 1 RFU) and no stutter filter to enable capture of noise and stutter values. Results of negative controls should not be used.
“Thus, the Calibration Files should be generated using as many single source profiles as possible, titrated to at least 5 input template masses ranging from the lowest mass typically processed in the laboratory to a maximum of 1ng, utilizing one amplification protocol and one injection time"1
“Where possible, Calibration Files should incorporate all laboratory processes such that intermediate levels of reproducibility are measured. For example, if the laboratory utilizes multiple extraction techniques, a portion of the Calibration Files should be treated with each extraction technique utilized. The Calibration Files should also capture a variety of dilution extracts post-qPCR. Calibration Files should be generated using multiple quantification, amplifications and capillary lot numbers.”1
Summary Checklist: When to Add More Samples
| Observation on Plot | Status | Action |
| Points are evenly distributed across all tested target masses - See Figure 1 | Adequate | Proceed with validation. |
| Large gaps exist between mass points - See Figure 2 | Inadequate | Add samples at the missing mass intervals. |
| Fitted line looks unstable or skewed by a few extreme outliers | Inadequate | Add more samples to dampen outlier leverage. |
| Sparse data at the low-template/stochastic boundary | Inadequate | Add low-mass samples to capture accurate dropout/noise trends. |
Evaluate the calibration plots before using the calibration file to analyze NOC in samples.
For additional details, see How to evaluate calibration plots FAQ and section 3 in the NOCIt Technical Manual.
Figure 1. Example of a plot with points evenly distributed across all tested target masses, including dense data in the low-template/stochastic boundary.
Figure 2. The plot has a gap in the spread and low density at the upper mass, indicating that additional samples are required to develop the model at this locus.
1 NOCIt software technical manual.