This repository is about predicting the anti-fungal peptide sequences. Along with predicting, this code also finds the Train & Test accuracy with K- Fold method and 5 folds.
Files required:
-
train.csv: This file has three columns, the first column contains ID (sequence number), the second column contains contain the label (-1 for negative and +1 for positive sample) and the third column contains the sequence of peptide in a single sting.
-
test.csv: This file has two columns, the first column contains ID and the second column contains the peptide sequence.
The model will train based on train.csv and predict the labels of the peptide sequences in test.csv
The output will be of two columns: first column containing the ID (sequence number) and the second column will contain the predicted label (-1 for negative and +1 for positive).
python3 final.py -o <output File> -type {aminoacid | dipepeptide | tripeptide} -model {SVC | RFC} { {-gamma | -n_estimators} <positive value> | -optimalParameters }
- Type can only be ‘aminoacid’, 'dipeptide' or ‘tripeptide’. These are the different ways of parsing the input peptide.
- Models:
- SVC: Support Vector Classifier
- RFC: Random Forest Classifier
- -gamma and -n_estimators are the parameters we pass into these models.
- -optimalParameters is a mode where it runs the model multiple times and plots the accuracy for the parameters. This way, we can find out the optimal gamma/n_estimators for a particular model.
- gamma can only be used if model is SVC.
- n_estimators can only be used if model is RFC.
- gamma & -n_estimators have to be positive values
- outputFile is preferred with .csv format
- In -optimalParameters, 3 inputs will be asked: start, end, step indexes for the parameters.
- If -optimalParameters mode is used, -gamma < val > cannot be used for obvious reasons.
Model: RFC
Type: Tripeptide
n_estimators: 700
CSV stored in: best1.csv
Code in: best1.py
python3 final.py -o best1.csv -type tripeptide -model RFC -n_estimators 700
Output exported in best1.csv
Avg Train Accuracy: 1.0
Avg Test Accuracy: 0.8901960784313726
Model: SVC
Type: Tripeptide
n_estimators: 15
CSV stored in: best2.csv
Code in: best2.py
python3 final.py -o best2.csv -type tripeptide -model SVC -gamma 15
Output exported in best2.csv
Avg Train Accuracy: 0.9859803921568627
Avg Test Accuracy: 0.8741176470588237
Model: SVC
Type: Dipeptide
n_estimators: 91
CSV stored in: best3.csv
Code in: best3.py
python3 final.py -o best3.csv -type dipeptide -model SVC -gamma 91
Output exported in best3.csv
Avg Train Accuracy: 0.9981372549019607
Avg Test Accuracy: 0.8631372549019607
python3 final.py -o a.csv -type dipeptide -model SVC -gamma 91
Output exported in a.csv
Avg Train Accuracy: 0.9981372549019607
Avg Test Accuracy: 0.8631372549019607
python3 final.py -o b.csv -type tripeptide -model RFC -n_estimators 700
Output exported in b.csv
Avg Train Accuracy: 1.0
Avg Test Accuracy: 0.8874509803921569
python3 final.py -o a.csv -type aminoacid -model SVC -optimalParameters
Enter start parameter: 40
Enter end parameter: 200
Enter step parameter: 30
Gamma : 40
Avg Train Accuracy: 0.9235294117647058
Avg Test Accuracy: 0.851764705882353
Gamma : 70
Avg Train Accuracy: 0.9575490196078432
Avg Test Accuracy: 0.8623529411764705
Gamma : 100
Avg Train Accuracy: 0.9748039215686275
Avg Test Accuracy: 0.8619607843137255
Gamma : 130
Avg Train Accuracy: 0.9858823529411765
Avg Test Accuracy: 0.8564705882352941
Gamma : 160
Avg Train Accuracy: 0.9920588235294119
Avg Test Accuracy: 0.8564705882352941
Gamma : 190
Avg Train Accuracy: 0.993921568627451
Avg Test Accuracy: 0.8525490196078431
python3 final.py -o c.csv -type tripeptide -model SVC -n_estimators 700
Invalid set of arguments. Please see README
>python3 final.py -o c.csv -type tripeptide -model SVC
Invalid length of arguments. Please see README