abdullah13521 / pyphewas

MASI Lab Implementation of PheWAS in Python

PyPheWAS

Repository for the pyPheWAS project. Full documentation at https://pyphewas.readthedocs.io/en/latest/

Developers

Cailey Kerley, PhD Candidate

Shikha Chaganti, PhD

Bennett Landman, PhD

Cite pyPheWAS

Kerley, C.I., Chaganti, S., Nguyen, T.Q. et al. pyPheWAS: A Phenome-Disease Association Tool for Electronic Medical Record Analysis. Neuroinform (2022). https://doi.org/10.1007/s12021-021-09553-4

Latest Release: pyPheWAS 4.1

4.1.1

• Minor Bug Fixes

4.1.0

• pyPheWAS Explorer updates
• New demographic variables added to synthetic dataset

4.0.5

• convertEventToAge includes new warning for calculated ages are negative
• small bugs fixed in maximizeControls, NoveltyAnalysis, and PubMedQuery tools

4.0.4

• createPhenotypeFile now supports more options for controlling case/control group curation
• Documentation updates

4.0.3

• Novelty Analysis tools: examine the relative literary novelty of disease-phecode pairings
• pyPheWAS Explorer: an interactive visualization of PheDAS experiments
• createGenotypeFile updated - now called createPhenotypeFile
• Minor bug fixes

Older Releases

pyPheWAS 3.2.0

• Configurable threshold for number of subjects required to run the regression on an individual PheCode
• All regressions are now fit with regularization (old scheme available with 'legacy' option)
• Minor changes to Manhattan plot
• PheCode/ProCode categories added to regression file
• Minor bug fixes

pyPheWAS 3.1.1

• New Analysis Type: ProWAS Tools
• New Plot Type: Volcano Plot (see pyPhewasPlot)
• maximizeControls now saves explicit Case/Control matches
• New PheCode category colors in plots are more distinguishable
• Improved command line tool argument handling
• Improved error handling
• Documentation overhaul
• API update
• Minor bug fixes

pyPheWAS 3.0.1

• Bug fixes including FDR & Bonferroni threshold calculations
• Header saved in feature matrices
• More file formats available for saving plots

pyPheWAS 3.0.0

• Support for both ICD 9 and ICD 10
• All 3 regression types (binary, count, & duration) optimized for big data
• pyPhewasPipeline: a streamlined combination of pyPhewasLookup, pyPhewasModel, and pyPhewasPlot
• Compatibility with Python 3
• Age matching now saves the explicit mapping between controls/cases in addition to the resulting group file
• Operation of the ICD censoring function matches the description in the documentation
• Minor bug fixes