eHarmonize

Contact: Alyssa Zhu alyssahz@usc.edu

ENIGMA Harmonization (eharmonize) is a python-based package that harmonizes provided data to included lifespan reference curves. The package is currently set up for outputs of the ENIGMA-DTI pipeline.

Please see wiki for additional details and information.

Current version (Release v0.0.0):

Installation

As this package has not been added to PyPI, please follow the instructions below carefully.

One-step process: pip install git+https://github.com/ahzhu/eharmonize.git

Two-step process:

1. To clone the repository: git clone https://github.com/ahzhu/eharmonize.git

2. After downloading or cloning (in the same directory that git clone was run in): pip install ./eharmonize

N.B. The above installation commands will require administrator/root privileges of the system. If you don't have administrator/root privileges, please try adding the --user flag.

Example: pip install --user ./eharmonize

Usages

Display available subcommands

eharmonize --help

Display flags/options for a specific subcommand

eharmonize [subcommand] --help

Use subcommand

eharmonize [subcommand] [--flags inputs]

Subcommands

harmonize-fa

The harmonize-fa subcommand takes in FA spreadsheets from the ENIGMA-DTI pipeline and exports FA values harmonized to the included reference curves.

The input CSV should have columns for Age and Sex. Optional columns include SITE and Dx. If SITE is not included, all data will be assumed to have come from a single site/scanner. If Dx is not included, all data will be assumed to have come from controls. In the Dx column, subjects should be denoted as either case or control. The harmonization parameters will be calculated based on the control group and then applied to the cases.

N.B. We currently do not support harmonization for case-only studies. If such data is detected, the program will self-quit.

Required Inputs:

• --incsv CSV File path of the CSV with the necessary covariates and FA measures
• --outdir DIR Directory to write out outputs to

Optional Inputs:

• --reference version Version number of desired reference [default: v0.1]
• --rerun If used, will overwrite previous outputs

Example usage:

eharmonize harmonize-fa --incsv enigma_FA_spreadsheet.csv --outdir /path/to/write/to eharmonize harmonize-fa --incsv enigma_FA_spreadsheet.csv --outdir /path/to/write/to --reference v0.0 --rerun

apply-harmonization

The apply-harmonization subcommand will apply an existing harmonization model to new data. The input CSV should follow the same requirements as in harmonize-fa. The --model and --log inputs are outputs from harmonize-fa.

Required Inputs:

• --incsv CSV File path of the CSV file with the necessary covariates and FA measures
• --model MODEL Model file as output by initial harmonization
• --log JSON JSON file from initial harmonization pant IDs to include
• --outdir DIR Directory to write out outputs to

Optional Inputs:

• --metric [default: FA]

Example usage:

eharmonize apply-harmonization --incsv enigma_FA_new_spreadsheet.csv --model --log --outdir eharmonize apply-harmonization --incsv enigma_FA_new_spreadsheet.csv --model --log --outdir --metric FA

harmonize-dti

In Development

Acknowledgments

We would like to thank the study coordinators and participants of the public datasets used to create the DTI reference curves.

• PING (Jernigan et al. 2016)
• HCP-Development (Somerville et al. 2018)
• ABCD (Hagler et al. 2019)
• SLIM (Qiu et al.)
• HCP (Van Essen et al. 2013)
• Cam-CAN (Shafto et al. 2014)
• PPMI (Marek et al. 2011)
• UK Biobank (Miller et al. 2016; Application Number 11559)
• OASIS3 (LaMontagne et al. 2019)
• ADNI3 (Weiner et al. 2016)