PPMI Parkinson's Progressive Markers Initiative

About

This repository has code that is used to process data from the PPMI dataset. There are no data files stored in this repository.

Setup the database

1. Apply for data access at PPMI- Download Data

2. After your application is approved and you can log in

   1. click on Download -> Study-Data.
   2. Click the link on the left hand side (bottom) where it says ALL
   3. Download ALL tabular data (csv format) and ALL documents and zip files [48.0 MB]

3. Clone this github repository

4. Save the downloaded files in the data_docs directory of this cloned repository. Extract any zip files if necessary.

5. Run the create_ppmi_database.py script which will create the ppmi database in the database directory

cd scripts/python
python scripts/create_ppmi_database.py 

Using the database

1. Install DB Browser for SQLite and use it to open the database that was created for you in the database directory.

2. DB Browser will give you information on all the tables in the database. You can browser the data by table. After you have extracted information into the data_docs directory (see Setup the database), you will get further information on what the columns represent by looking at the files in the docs directory.

   • Page_Descriptions.csv A general idea of what the tables are named
   • Data_Dictionary.csv A general idea of what the columns in the table are named
   • Code_List.csv A code book that tells us how certain variables are coded
   • Derived Variables Definition A document that tells us how to create derived variables like UPDRS scores etc.

3. Use SQL scripts to extract information from the database.

WITH A AS ( /* Calculated Part III UPDRS Score for everyone */SELECT PATNO, EVENT_ID , NP3SPCH + NP3FACXP + NP3RIGN + NP3RIGRU + NP3RIGLU + PN3RIGRL + NP3RIGLL + NP3FTAPR + NP3FTAPL + NP3HMOVR + NP3HMOVL + NP3PRSPR + NP3PRSPL + NP3TTAPR + NP3TTAPL + NP3LGAGR + NP3LGAGL + NP3RISNG + NP3GAIT + NP3FRZGT + NP3PSTBL + NP3POSTR + NP3BRADY + NP3PTRMR +
NP3PTRML + NP3KTRMR + NP3KTRML + NP3RTARU + NP3RTALU + NP3RTARL + NP3RTALL + NP3RTALJ + NP3RTCON as UPDRS_SCORE
FROM NUPDRS3 WHERE PAG_NAME ="NUPDRS3"ORDER BY PATNO)
/* Extract UPDRS scores at Baseline and sort in descending order */SELECT*from A WHERE EVENT_ID="BL"ORDER BY UPDRS_SCORE DESC

The output in SQLite Browser looks something like this for the first few rows. It can be exported to a csv file for further analysis.

PATNO	EVENT_ID	UPDRS_SCORE
42142	BL	72
42435	BL	72
40763	BL	71
42439	BL	67
40751	BL	66
42312	BL	62
50496	BL	61
51652	BL	61
41489	BL	60

You can find more complex sql scripts in the scripts directory