autoBioSeqpy: A Deep Learning Tool for the Classification of Biological Sequences

Area of Science:

• Bioinformatics
• Computational Biology
• Machine Learning

Background:

• Deep learning models are powerful for analyzing diverse data, including biomedical sequences.
• Current deep learning toolkits (e.g., TensorFlow, PyTorch) often require specialized implementations, limiting transferability across research projects.
• Biological sequence classification remains a computationally intensive task requiring accessible tools.

Purpose of the Study:

• To present autoBioSeqpy, a novel deep learning tool designed for simplified biological sequence classification.
• To offer a user-friendly command-line interface for automated deep learning model execution.
• To provide adaptable model templates for enhanced usability in biological sequence analysis.

Main Methods:

• autoBioSeqpy automates key deep learning steps: data input, parameter initialization, sequence encoding, model training, and evaluation.
• The tool utilizes a command-line interface, requiring minimal user input beyond data preparation.
• It incorporates customizable and ready-to-use model templates for diverse applications.

Main Results:

• autoBioSeqpy successfully applied to three distinct biological sequence classification tasks: Type III secreted protein prediction, protein subcellular localization, and CRISPR/Cas9 sgRNA activity prediction.
• Demonstrated simplicity and efficiency in executing deep learning pipelines for biological sequence analysis.
• Validated the tool's capability to handle various biological sequence datasets and classification problems.

Conclusions:

• autoBioSeqpy offers a streamlined and accessible approach to applying deep learning for biological sequence classification.
• The tool enhances research efficiency by automating complex modeling steps and providing adaptable templates.
• It represents a valuable resource for researchers in bioinformatics and computational biology seeking to leverage deep learning.