Skip to main content

AI assistant and notebook executable code blocks

Our AI assistant is a powerful tool designed to help life scientists quickly analyze and visualize their data. This feature combines the ease of natural language, the power of scientific Python libraries such as RDKit, and interactive data visualization tools such as Plotly. By leveraging our AI assistant, researchers can accelerate their data processing, explore data efficiently, and easily incorporate advanced Python-based analyses into their research.

Example use cases​

Our AI assistant is particularly useful for tasks such as:

  • Quickly generating code for common data processing and analysis tasks.
  • Interactively exploring data with plotting libraries such as Bokeh and Plotly.
  • Creating custom figures of experimental data.
  • Performing statistical analyses on datasets.
  • Generating boilerplate code for more complex analyses which users can further customize.
  • Enable researchers who have limited familiarity with Python to access its power.

As an example, the demo video below illustrates how our AI assistant can be used to visualize the clustering of single-cell sequencing data, utilizing Python and seaborn.

Starting the AI assistant​

There are two primary ways to access the AI assistant:

  • Data-driven queries
    1. Navigate to the database you'd like to analyze.
    2. Select the rows and columns of the database that you'd like to analyze.
    3. Right-click on these rows and columns and select "Start AI chat" in the context menu. This will open our AI assistant at the right side of your screen.
  • General queries
    1. Click the chat button located at the lower right corner of our platform.

Features and functionality​

Python code generation​

Our AI assistant generates Python code based on natural language queries. It can handle a wide range of tasks, including:

  • Simple calculations
  • Complex data analysis
  • Interactive data visualizations
  • Statistical analyses

Data-aware code generation​

When accessed from selected rows and columns, our AI assistant automatically receives information about the selected data. This allows our assistant to generate Python code for analyzing and visualizing your data.

Data retrieval​

Our AI assistant works seamlessly with your databases. Specifically, the code generated by our assistant retrieves data using the deeporigin.data_hub module of our Python client.

Extensive scientific libraries​

Our assistant is pre-loaded with many biology and data science capabilities from Python packages such as:

Code execution​

Users can easily execute the code generated by our assistant:

  1. After running a query, click the "Execute code" button located at the top right of the block with the generated code.
  2. View the output immediately below the generated code.

Code modification​

Users can modify the code generated by our AI assistant, allowing for customization such as changing the color of a chart or the title of an axis.

Integration with notebooks​

To save analyses and visualizations, users can embed the code generated by our assistant and their results into their notebooks:

  1. After executing code generated by our assistant, click the "Insert into notebook" button at the top right of the code block.
  2. The code and its results will then be inserted into your notebook at the location of your cursor, ensuring a complete record of your analysis.
  • Be specific: The more specific your query, the more accurate and useful the generated code will be.
  • Review and fine-tune: Always review the generated code and modify it as necessary. The AI assistant often provides excellent starting points; your expertise is crucial for ensuring the analysis meets your specific needs.
  • Iterate: Use our AI assistant iteratively. Start with a basic query, refine your query based on the initial results, and continue to ask our AI assistant questions to further fine-tune it.