Protocol for vision transformer-based evaluation of drug potency using images processed by an optimized Sobel...

Area of Science:

• Biotechnology
• Computational Biology
• Drug Discovery

Background:

• Traditional anticancer drug screening is often time-consuming, labor-intensive, and expensive.
• There is a need for more efficient and cost-effective methods to evaluate drug efficacy.

Purpose of the Study:

• To present a novel protocol for label-free, high-throughput drug efficacy assessment.
• To utilize deep learning, specifically a vision transformer and Conv2D, for predicting drug potency.

Main Methods:

• The protocol involves cell culture, drug treatment, and data collection.
• Deep learning models were developed using vision transformer and Conv2D architectures.
• Data preprocessing steps were detailed for model training.

Main Results:

• The developed deep learning models can predict drug potency with high throughput.
• The label-free approach eliminates the need for costly chemical reactions.
• The protocol is adaptable for screening chemicals impacting cell density or morphology.

Conclusions:

• This novel protocol offers a faster, cheaper, and more efficient alternative for anticancer drug screening.
• The methodology can be extended to screen various chemical compounds affecting cellular characteristics.
• The integration of deep learning with cell imaging accelerates the drug discovery pipeline.