"Evaluation of Trainable Protein Representation Methods: A Quantitative Review" is an evalutation and review study on trainable protein representations.
-
The study is an investigation of the available trainable protein representation methods.
-
Aiming to evaluate how much each representation model captures different facets of functional information, we constructed and applied benchmarks based on;
- The detection of semantic similarities between proteins (a.k.a Semantic similarity inference)
- Ontology-based protein function prediction
- Drug-target protein family classification
-
The study is summarized in the figure below;
-
ProtBench is tool that can be used to evaluate any protein representation model. The tool can be found under generalized_representation_benchmark folder (Still under development)
-
Benchmark_Study folder includes codes used for benchmarks presented in the paper.
-
go_prediction folder is under Benchmark_Study and includes codes used for "Ontology-based protein function prediction" task.
-
embedding_similarity folder is under Benchmark_Study and includes codes used for "Semantic similarity inference" task.
-
protein_family_prediction folder is under Benchmark_Study and includes codes used for "Drug-target protein family classification" task.
-
preprocess folder is under Benchmark_Study and includes codes used for data preprocessing for "Ontology-based protein function prediction" and "Semantic similarity inference" tasks.
-
The large files used for the benchmark is shared on GDrive and the main directory structure is shown below.
The data used in ProtBench was located under "ProtBench" folder. The data belongs to benchmark study was located under the "Benchmark_IO_data" folder. Each task and shared data has its own folder. The directory structure is self-explanatory and standard for all tasks, hence some of the folders might be empty based on task for now.
Other than that the reusable 11 precalculated reusable protein representation vectors for human proteins can be found under "Shared_Data/Representations/representations_vectors/representation_vector_dataframes" folder for further use.
-ProtBench
-Benchmark_IO_data
|
|---GO_prediction
|
|---preprocess
|---input_data
|---results
|---main_results
|---detailed_results
|---Embedding_Similarity
|
|---preprocess
|---input_data
|---results
|---main_results
|---detailed_results
|---Protein_Family_Prediction
|
|---preprocess
|---input_data
|---results
|---main_results
|---detailed_results
|---Shared_Data
Copyright (C)
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.