Newport iServer MicroServer iTHX-M Operator's Manual page 353

CONTAMINATION QC CHECK
Notice that even with RS < threshold the genotype call still could be
no call since the Null model may be the most likely model.
For more information, see
page 347.
Adjustment of thresholds
Increasing the threshold can reduce the number of no calls, but may
also reduce the confidence of the calls.
Decreasing the threshold can increase the confidence of the calls, but
may also increase the number of no calls.
Displaying Rank Scores in the DM Scatter Plot
The DM Scatter Plot (see
page 137) displays four different values (rank scores for AA, AB, BB,
and null genotype calls) on a two-dimensional graph. The following
transformations are applied to the data to accomplish this and allow
comparisons of genotype calls for different samples:
• The first transformation accentuates the areas of interests in the
rank score. The scale is 0-1, but genotypes are only called for rank
scores less than 0.25. The first transformation is designed to
magnify the important small rank score differences and lessen
large rank score differences which aren't significant.
• A second transformation is used to convert to a range 0-1.
• The final transform maps the values to a two dimensional plot.
Because of the importance of detecting mixed or contaminated
samples, GTYPE 4.1 has implemented an additional algorithm to
supplement the Dynamic Model algorithm to help with sample
contamination detection. The MPAM calling algorithm, used to make
genotype calls for the GeneChip 10K mapping arrays, is employed in
this analysis to make genotype calls on a subset of ~ 8,000 SNPs on
DM mapping arrays. These SNPs were chosen based upon robust
Appendix C | Dynamic Model Mapping Algorithm
One-sided Wilcoxon's Signed Rank Test, on
Chapter 5, Mapping Analysis Window, on
345