Problem solved

Spatial transcriptomics researchers face a fragmented landscape of analysis tools, each with its own interface, dependencies, and parameter requirements. This leads to steep learning curves, inefficient workflows, and barriers to multi-method analyses. ChatSpatial solves this by providing a unified, natural language interface that abstracts away complexity, enabling researchers to execute sophisticated analyses without needing to recall exact command syntax or method-specific quirks. It democratizes access to advanced spatial analysis by letting users simply describe what they want to do in plain English.

Use cases

• Performing cell type annotation on spatial transcriptomics datasets using natural language commands
• Running deconvolution algorithms to estimate cell type proportions in spots or regions
• Computing spatial autocorrelation statistics (e.g., Moran's I) to identify gene expression patterns
• Generating spatial visualizations like heatmaps or scatter plots on tissue images
• Combining multiple analysis steps into a pipeline by simply describing the desired outcome to an AI assistant
• Exploratory analysis where a researcher iteratively asks questions about the data and gets immediate results

Audience

• Developers
• Data Scientists
• Researchers

Pros

• Integrates 60+ established methods, reducing tool fragmentation
• Conversational AI interface lowers the barrier for non-programmers
• Open-source and self-hosted, offering full data privacy and customization
• Schema-enforced orchestration prevents common parameter errors
• Docker support ensures reproducible and easy deployment

Cons

• Limited to spatial transcriptomics, not general-purpose bioinformatics
• Requires technical setup (Python environment, possibly Docker) to run locally
• Dependency on external packages may lead to large installations and potential version conflicts