Tutorial

Before taking this short tutorial, PowQ should be installed in your system.

The following examples are intended briefly to illustrate the program functionality. Indeed, only a small number of simulations will be used in this tutorial, generally insufficient for a real power study. The comparative analysis example should take about 5 minutes and the multipoint analysis should take about 20 minutes.

Comparative analysis

For an example of comparative analysis download the pedigree files PedigreeA.ped and PedigreeB.ped. The two pedigree structures differ in the type of sampling unit. PedigreeB.ped contains two unrelated nuclear families, while in PedigreeA.ped the two nuclear families share a common ancestor couple. The number of individuals for whom the phenotypic and genotypic information are available is the same in both the sampling scheme (eight, all having value “2” in the pedigree file selection flag and represented with blackened boxes). For a fast power comparison of the two pedigrees download the Map0.map file, which contains a single marker coincident with the QTL. Consult the input files section for some explanations about the files format.

Start PowQ and specify the input files in the “Input data” panel: PedigreeA.ped for “Ped file (A)”, PedigreeB.ped for “Ped file (B)”, and Map0.map for “Map file”.

To select the files use the “Browse” buttons. PowQ will look in the current folder for files with the extension “.ped” when selecting the pedigree files and for files with the extension “.map” when selecting the map file. In the same panel it is possible to specify the LOD score thresholds that will be used in the analysis. For this example we will use 2.0 and 3.0 as suggestive and significant LOD-score respectively, by writing these values in the “Suggestive LOD” and “Significant LOD” fields. We will test 300 pedigrees having a structure defined in the two pedigree files - corresponding to 300 extended families and 600 nuclear families (300 * 2) – by filling the field “# of replicates” with 300. In order to test the two datasets power using 100 simulation cycles, fill the field “# of simulations” with the value 100. For any explanation about the parameters defined in the input data panel, please read to the input data section.

The genetic model will be defined in the “Genetic model” panel, accessible by pressing the button placed on the top of the PowQ window.

We will define a quantitative phenotype with narrow heritability of 60%, due to a QTL accounting for the 30% of variability and polygenic effects accounting for the remaining 30%. This model can be specified by writing the value “1.850” in the “Displacement” field and “0.430” in the “Residual h2” field, and updating the parameters by pressing the “”Evaluate” button. These parameters are explained in the input data section. Further details about the simulation process are described in the implementation details section.

To run the power study press the “Run” button on the top of the PowQ window. In the “Run” panel press the “Start” button. A progress bar on the bottom of the panel will show the elapsed and remaining time. The other sub-panels can be explored during the simulations and will show the simulations outcome in real time.

At the end of the simulation process the “LOD Freqs (A)” and “LOD Freqs (B)” sub-panel will show the observed power for the suggestive, significant and highly significant thresholds for the two sampling schemes. By comparing the power of the two samples for the significant threshold, selecting those nuclear families linked by a common ancestor couple would provide more than the double of the power for the selected model (95% vs 44%).

Note: by evaluating 1,000 simulation cycles the power ratio is about the same (91% vs 41%).

Multipoint analysis

This second example needs more computational time, since the benefits of a multipoint analysis become evident only after a larger number of simulations. For an example of multipoint analysis download the pedigree file PedigreeB.ped (used also in the previous example). PedigreeB.ped describe two nuclear families. The analysis will be carried out on all pedigree members, having value “2” in the pedigree file selection flag and represented with blackened boxes. Start PowQ and specify the input files in the “Input data” panel: PedigreeB.ped for “Ped file (A)”.

For a fast power comparison of the single-point and multipoint power computation download the Map2.map file, which contains two markers flanking the QTL.

To select the files use the “Browse” buttons. PowQ will look in the current folder for files with the extension “.ped” when selecting the pedigree files and for files with the extension “.map” when selecting the map file. Consult the input files section for some explanations about the files format.

In the same panel it is possible to specify the LOD score thresholds that will be used in the analysis. For this example we will use 2.0 and 3.0 as suggestive and significant LOD-score respectively, by writing these values in the “Suggestive LOD” and “Significant LOD” fields. We will test 700 pedigrees having a structure defined in the pedigree file - corresponding to 1,400 nuclear families – by filling the field “# of replicates” with 700. In order to test the two dataset with 500 simulation cycles, fill the field “# of simulations” with the value 500. For any explanation about the parameters defined in the input data panel, please read to the input data section.

The genetic model will be defined in the “Genetic model” panel, accessible by pressing the button placed on the top of the PowQ window.

We will define a quantitative phenotype with narrow heritability of 70%, due to a QTL accounting for the 40% of variability and polygenic effects accounting for the remaining 30%. This model can be specified by writing the value “2.310” in the “Displacement” field and “0.500” in the “Residual h2” field, and updating the parameters by pressing the “Evaluate” button. These parameters are explained in the input data section. Further details about the simulation process are described in the implementation details section.

To run the power study press the “Run” button on the top of the PowQ window. In the “Run” panel press the “Start” button. A progress bar on the bottom of the panel will show the elapsed and remaining time. The other sub-panels can be explored during the simulations and will show the simulations outcome in real time (alternatively have a coffee, on smaller computers evaluation of all the simulations can be slow).

At the end of the simulation process the “LOD Freqs (A)” sub-panel will show the observed power for the suggestive, significant and highly significant thresholds. The single-point power for the significant threshold of 3.0, averaged between the values obtained by the single markers, is about 70%, while the power in the entire region is more than the 80%. This latter estimation considers as significant each simulation showing a significant LOD score at least at one of the two markers.

Note: by evaluating 1,000 simulation cycles the power ratio is about the same.