A Blood-Test Based Predictive Model for the Cognitive Decline in Probable Alzheimer’s Disease Patients
This repository include the code and result of prediction model building, training, and fine-tuning. The model is intended to predict cognitive decline in probable AD patients using the exRNA level of a specific gene, PHGDH.
The cohort we used to train the predictive model comes from Zhong lab’s previous study, Presymptomatic Increase of an Extracellular RNA in Blood Plasma Associates with the Development of Alzheimer’s Disease. Zhangming Yan, Zixu Zhou, Qiuyang Wu, Zhen Bouman Chen, Edward H. Koo, Sheng Zhong. Current Biology, 2020, 30:1771–1782. The TPM_normalized data is publicly available on GEO: GSE136243.
.
├── data <- includes the sample formatted longitudinal dataset
│ └── ADRC_PHGDH_with_cogscore_noAD.csv
├── README.md
├── result <- includes model performance, graph, etc.
│ ├── algorithm_selection <- performances of the model trained with three versions of PT checker
│ │ ├── PT_version1.csv
│ │ ├── PT_version2.csv
│ │ └── PT_version3.csv
│ ├── modeling <- illustrations of the model, and graphs for variable comparison
│ └── parameter_fine_tuning.csv <- accuracy, PPV, NPV, and other specific information of models trained with 11664 candidate parameter combinations
└── script <- includes scripts for the prediction model algorithm and further analysis
├── algorithm <- scripts for the prediction model algorithm
│ ├── CF.R <- cognitive function checker, label the cognitive score
│ ├── prediction_model_conditionalCF.R <- function that combine two labeling functions, and evaluate the model performance
│ ├── prediction_model_currentCF.R
│ ├── prediction_model_current+futureCF.R
│ ├── PT_version1.R <- predictive test checker, label the exRNA PHGDH level
│ ├── PT_version2.R
│ └── PT_version3.R
├── modeling_result_illustartion.R <- code for modeling result illustration
├── parameter_fine_tuning.R <- code for setting loops for parameter fine-tuning
├── PT_algorithm_selection.R <- code that includes the candidate parameter sets for three versions of PT checkers
├── set_individual_coordinates.R <- an environment code referenced from online resources, allowing users to set individual coordinates for each facet in ggplot2
└── storing_variables.R <- setup variables for information storing during the model training process
Last update: Aug 18, 2022