University of Pittsburgh researchers are developing novel software, spatiAlytica, an AI-powered user-friendly system for analysis of spatial transcriptomics (ST) data. This novel software is designed for use by biologists without specialist coding or bioinformatics knowledge to interact with ST data by asking questions or giving instructions. Analysis can be performed in real time, allowing users to visualize data, test hypothesis, or gain insight into complex data and could improve workflow for biologists and accelerate discovery in basic and translational biomedical research.
An AI-powered software, spatiAlytica, is being developed to assist biologists with analysis of ST data. The software does not require specialist coding or bioinformatics knowledge. Users can pose complex questions in natural language to the software in the form of a chatbot with spatiAlytica generating and executing the code required to analyze the data. This approach will streamline biological research.
Description
ST, where gene expression is determined for spatial locations in a tissue sample, is a powerful tool used by biologists to understand the role of tissue organization (i.e., the microenvironment) in gene expression. It is particularly useful in cancer research and has uses in drug discovery through identification of therapeutic targets. Currently, detailed analysis of ST data can be complex and inefficient. spatiAlytica is designed to make analysis of data more intuitive and accessible for biologists, reducing the need for coding knowledge, and allowing biologists to explore data at their own pace. Additionally, spatiAlytica can be used as an educational tool to inspire and train the next generation of bioinformaticians and data scientists.
Applications
• Cancer research
• Neurobiology research
• Biological research
Advantages
Current analysis of ST data is complex and time-consuming requiring biologists to collaborate with bioinformaticians with coding skills. Complications including scheduling conflicts, the need for follow up meetings and different knowledge bases between specialists can slow down research or lead to inaccurate analysis.
spatiAlytica is designed to work as an AI informatician. Without coding knowledge, ST datasets in multiple formats can be loaded and analyzed by users directly, asking a chatbot questions in over 80 spoken languages. These questions are converted to Python code and data analysis is performed in real-time. This approach allows for intuitive data exploration and analysis without the need for bioinformaticians, speeding up research workflows, enhancing collaboration with specialists (i.e., bioinformaticians can focus on novel algorithms), and improving output for resource-poor researchers who may struggle to access specialists. spatiAlytica is designed to securely handle sensitive data making it suitable for clinical and translational research, is highly adaptable and scalable, and can be readily incorporated into existing workflows.
Invention Readiness
Currently in development. A cancer-specific large-language-model that combines advanced natural language processing capabilities that works with an interactive visual interface built on the Napari platform has been developed. Work is now ongoing to enhance functionality and performance.
IP Status
Patent Pending
