# Windows Installation¶

The following steps will help you install pyGenClean on a Windows machine. It has been tested on both Windows XP and Windows 7.

## Requirements¶

The following softwares and packages are required for pyGenClean:

1. Python 2.7
3. numpy (version 1.6.2 or latest)
4. matplotlib (version 1.2.0 or latest)
5. scipy (version 0.11.0 or latest)
6. scikit-learn (version 0.12.1 or latest)
7. Jinja2 (version 2.7.3 or latest)

Note

All the requirements will be installed along with the main pyGenClean module.

Warning

The Plink software needs to be in the PATH (or in the current working directory). In other words, you should be able to type plink at the command line.

## Installation¶

The easiest way to install Python on windows is by using Miniconda.

### Miniconda¶

To create a new virtual environment, perform the following command:



#### Installing pyGenClean¶

To install pyGenClean, only perform the following command:



## Testing the Algorithm¶

To test the algorithm, download the test data from http://statgen.org/downloads/pygenclean/ and the HapMap reference populations (build 37). Create a directory on your Desktop named pyGenClean_test, and extract the two archive into it. You should have the following directory structure:

Desktop\
pyGenClean_test_data\
check_ethnicity_HapMap_ref_pops_b37\
hapmap_CEU_r23a_filtered_b37.bed
hapmap_CEU_r23a_filtered_b37.bim
hapmap_CEU_r23a_filtered_b37.fam
hapmap_YRI_r23a_filtered_b37.bed
hapmap_YRI_r23a_filtered_b37.bim
hapmap_YRI_r23a_filtered_b37.fam
hapmap_JPT_CHB_r23a_filtered_b37.bed
hapmap_JPT_CHB_r23a_filtered_b37.bim
hapmap_JPT_CHB_r23a_filtered_b37.fam


Open the command prompt and navigate to the newly created directory, and created an new text file using notepad:

> cd Desktop\pyGenClean_test


Insert the following code in the file:

  1 2 3 4 5 6 7 8 9 10 [1] script = check_ethnicity ceu-bfile = check_ethnicity_HapMap_ref_pops_b37/hapmap_CEU_r23a_filtered_b37 yri-bfile = check_ethnicity_HapMap_ref_pops_b37/hapmap_YRI_r23a_filtered_b37 jpt-chb-bfile = check_ethnicity_HapMap_ref_pops_b37/hapmap_JPT_CHB_r23a_filtered_b37 nb-components = 2 multiplier = 1 [2] script = sex_check 

Finally, run the following command:

> run_pyGenClean ^
--conf conf.ini ^


### Results¶

Valuable information will be shown on the command prompt. Once the program has finished, the results are in the new directory data_clean_up.date_time where date is the current date, and time is the time when the program started.

Here are the new directory structure, with only the files you might be interested in:

• data_clean_up.data_time\
• pyGenClean.log
• automatic_report.tex
• excluded_markers.txt
• excluded_samples.txt
• 1_check_ethnicity\
• ethnicity.before.png
• ethnicity.outliers.png
• ethnicity.outliers
• ethnicity.population_file_outliers
• 2_sex_check\
• sexcheck.list_problem_sex

The file pyGenClean.log contain the information that was displayed in the console. The file automatic_report.txt contain the automatic report generated by pyGenClean. The files excluded_markers.txt and excluded_samples.txt contains the list of markers and samples, respectively, that were excluded from the dataset (with the reason).

#### 1_check_ethnicity/¶

The first image in the first directory (ethnicity.before.png) shows the MDS values for each sample before outlier detection. The second image (ethnicity.outliers.png) shows the outliers that should be removed for further analysis. Finally, the file ethnicity.outliers include a list of samples that should be removed for further analysis. The total number of outliers for this test should be exactly 63. For more information about the results of this module, refer to Section Ethnicity Module.

ethnicity.before.png

ethnicity.outliers.png

#### 2_sex_check/¶

In the second directory, there should be a file containing the list of samples with gender problem. There should be exactly 4 samples with gender problem. For more information about this module, refer to Section Sex Check Module.

If you want to compare your results with the expected ones, just download the files in the archive pyGenClean_expected_results.tar.bz2, available through http://statgen.org/downloads/pygenclean/. They were generated using Fedora 18 (64 bits) in about 20 minutes. You should at least compare the following files:

• 1_check_ethnicity
• ethnicity.outliers
• ethnicity.population_file_outliers
• All the figures (they might be mirrored).
• 2_sex_check
• sexcheck.list_problem_sex
• sexcheck.list_problem_sex_ids

### Automatic report¶

If LaTeX is installed, you can perform the following commands to compile the automatic report into a PDF file.

$pdflatex automatic_report.tex$ pdflatex automatic_report.tex
\$ pdflatex automatic_report.tex


The following PDF report will be generated.