Full project with Conv1D, tutorials and README

README.md

@@ -1,8 +1,42 @@
-# cafa5
+# Protein function prediction
 
-# install
+A python package for the prediction of protein functions as GO terms to participate in the [CAFA-5](https://www.kaggle.com/competitions/cafa-5-protein-function-prediction/overview) competition. 
 
-``` console
+The development and testing of the functionalities were made using python3.8, on a linux ─ubuntu 20.4─ environment.
+
+# Install
+
+(optional) Create a virtual environment to install the package.
+
+```
 $ mkdir -p venv && python3 -m venv ./venv && source venv/bin/activate
+```
+
+And install it.
+
+``` 
 $ python3 -m pip install .
 ```
+
+# Use
+
+Enter a subshell in the virtual environment with dependencies available:
+
+```
+$ source venv/bin/activate
+```
+
+and run python tasks from inside the subshell.
+
+## Usage tutorials
+
+- [Tutorial1](analyses/01_working_with_proteins.ipynb): Working with proteins and and protein structures.
+- [Tutorial2](analyses/02_get_go_terms_train_set.ipynb): Computing GO terms on the training set.
+- [Tutorial3](analyses/03_get_candidate_terms_test_set.ipynb): Computing candidate GO terms for model training.
+- [Tutorial4](analyses/04_get_ordered_candidate_terms_by_ia.ipynb): Order candidate terms by information accrued.
+- [Tutorial5](analyses/05_create_model_for_top_scored_term.ipynb): Creating a single model and benchmarking its accuracy.
+- [Tutorial6](analyses/06_run_models_massive.ipynb): Massive training of models.
+- [Tutorial7](analyses/07_massive_prediction.ipynb): Massive prediction of protein functions.
+
+
+