For each of the models we generated and compared, random test sets combined randomly against random test sets that had their values combined by the pairwise strategy.
The ERAP1 models we generated exhibit structural characteristics that appear highly relevant to the peptide library screen results.
We sought to validate the models we generated in two ways.
For both these models we generated 22 samples from a single population with constant population size.
To increase the number of CCDS1 models, we generated a novel murine model of CCDS1 obtained by ubiquitous deletion of 5 7 exons in the Slc6a8 gene.
Based on the LUR models, we generated a predicted spatial surface for annual average concentrations for each pollutant in a GIS with a resolution of 10 m.
Conservatively assuming small effects, the models we generated predicted that power would be above 0.9 if the total cross-site sample consisted of 450 or more cases.
The full atomic models we generated for the Azoarcus and Twort molecules contain full atomic detail for the loop regions that are missing in the crystal structure.
To determine the variables most consistently retained in the prediction models, we generated the 10 "best models," based on the proportion of variance explained, by calculating all possible logistic regression models retaining 10 variables.
To generate a measurement of the specificity of reads that were found to align to gene models, we generated a synthetic set of reads that were the reverse, but not the reverse compliment, of all observed sequences reads (entire dataset).
