changelog.txt

1. change repo name to 5UTR_Optimizer
2. add parameter options and restructure buildModel_final.R
3. create new R script evalModel_final.R for evaluation different choice of ML model
4. rename evolutionDesign_Ribo.R to evolutionDesign.R, and add corresponding parameters options
5. modify FeatureCommons.py, FeatureExtraction_final.py, FeatureExtraction_singleInput.py, finalFormat_3k_synthetic_seqs.py to be compatible for python 3
6. add environment.yml for all dependencies
7. modified readme with updated command example and software version
8. deleted unneccessary old scripts

Feb 27 2021
1. implement glmnet/rpart/svm and generate plot pdf for spearcor, R-square/normliazed version, feature importance
2. update readme to put evalmodel before the build model
3. python script plot_eval.py to plot excel file data generated by evalModel_final.R
4. output excel contains six tabs:
    - TE_result: the evaluation result for predicing TE for each model in 10 fold CVs  , sp_cor is spearman correlation, rsq is R-squared statstics
    - RNA_result:  the evaluation result for predicing RNA-seq RPKM for each model in 10 fold CVs  , sp_cor is spearman correlation, rsq is R-squared statstics
    - TE_features_accuracy: random-forest feature importance value based on prediction accuracy for TE
    - TE_features_impurity: random-forest feature importance value based on node impurity for TE
    - RNA_features_accuracy: random-forest feature importance value based on prediction accuracy for RNA-seq RPKM
    - RNA_features_impurity: random-forest feature importance value based on node impurity for RNA-seq RPKM