https://github.com/corybrunson/aatd-score

build risk scores for genetic abnormalities associated with alpha-1 antitrypsin deficiency

https://github.com/corybrunson/aatd-score

Science Score: 33.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
  • .zenodo.json file
  • DOI references
    Found 1 DOI reference(s) in README
  • Academic publication links
    Links to: ncbi.nlm.nih.gov
  • Committers with academic emails
    1 of 1 committers (100.0%) from academic institutions
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (11.6%) to scientific vocabulary
Last synced: 11 months ago · JSON representation

Repository

build risk scores for genetic abnormalities associated with alpha-1 antitrypsin deficiency

Basic Info
  • Host: GitHub
  • Owner: corybrunson
  • Language: R
  • Default Branch: main
  • Size: 125 KB
Statistics
  • Stars: 0
  • Watchers: 2
  • Forks: 0
  • Open Issues: 1
  • Releases: 0
Created about 4 years ago · Last pushed over 2 years ago
Metadata Files
Readme

README.md

Risk scores for severe AATD

This repo contains code used in the study by Riley, Brunson, Eydgahi, Brantly, and Lascano.[^1]

Objective

The aim of the study was to design an interpretable clinical decision support tool to inform decision to screen for abnormal AAT genotype. The modeling goal was to predict abnormal genotype from history of hepatic disease, respiratory disease, and (optionally) smoking.

Data source

This project relied on data collected and provided by the National Detection Program based at the University of Florida and GeneAidyx.

Software dependencies

The project used Tidyverse for general data science and Tidymodels for machine learning. Additional risk models were built using FasterRisk.

Most scripts source project-wide settings from 'code/settings.r'. Due to FasterRisk being implemented in Python and not accessible through Tidymodels, separate scripts were prepared to build and analyze these models.

Model selection

The first round of analysis (the scripts '1-*.r') compared the performance of several families of predictive models on a subset of the data, using a simple train--test partition without hyperparameter optimization. The goal was to inform decisions about a cross-validated modeling effort using the complete data set. Specifically, we used this preliminary analysis to (1) select predictors to consider, (2) decide what interpretable model families to use, and (3) how much predictive performance might be weakened by using interpretable rather than non-interpretable models.

Model optimization

The second round of analysis (the scripts '2-*.r') conducted a stratified 6-fold cross-validated comparison of models selected based on the results of the first round: (1) as predictors, either lung and liver history alone or together with either gender or smoking history; and (2) as model families, logistic regression, random forest, nearest neighbor, and FasterRisk, together with a stand-in for the current guidance to screen patients with COPD. This round used a majority training set of the data to determine optimal hyperparameter settings, with the remaining testing set held out for evaluation in the third round.

Model evaluation

The third and final round of analysis (the scripts '3-*.r') fitted the optimized model families and specifications to the full training set from the second round and evaluated them on the testing set.

Contact

Please email Cory Brunson with any question about the analysis and code.

[^1]: Riley EL, Brunson JC, Eydgahi S, Brantly ML, Lascano JE. Development of a risk score to increase detection of severe alpha-1 antitrypsin deficiency. ERJ Open Res. 2023 Sep 18;9(5):00302-2023. doi: 10.1183/23120541.00302-2023. PMID: 37727673; PMCID: PMC10505949.

Owner

  • Name: Cory Brunson
  • Login: corybrunson
  • Kind: user
  • Location: Gainesville, FL
  • Company: @LaboratoryForSystemsMedicine

Mathematician by training, data scientist by testing. Relatively new to pretty much everything.

GitHub Events

Total
Last Year

Committers

Last synced: about 1 year ago

All Time
  • Total Commits: 61
  • Total Committers: 1
  • Avg Commits per committer: 61.0
  • Development Distribution Score (DDS): 0.0
Past Year
  • Commits: 0
  • Committers: 0
  • Avg Commits per committer: 0.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
Jason Cory Brunson j****n@m****u 61
Committer Domains (Top 20 + Academic)

Issues and Pull Requests

Last synced: over 1 year ago

All Time
  • Total issues: 3
  • Total pull requests: 0
  • Average time to close issues: about 19 hours
  • Average time to close pull requests: N/A
  • Total issue authors: 1
  • Total pull request authors: 0
  • Average comments per issue: 0.0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Top Authors
Issue Authors
  • corybrunson (3)
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels