No. You can interact with Purna entirely through natural language. However, if you’re familiar with Python, you can ask Purna to write and run custom scripts directly in your browser using Pyodide with pandas, matplotlib, and seaborn.
What is the difference between Agent Mode and Case Mode?
Agent Mode is a chat-first interface for general genomics questions and database searches — no patient context needed.Case Mode is for managing patient cases with variant data, phenotypes, and clinical findings. Toggle between modes using the buttons in the top-right corner.
What file format is required for uploading variants?
Purna requires VCF (Variant Call Format) files. Upload during case creation by dragging and dropping or using the Upload VCF button.
What happens to my data?
Your data is securely stored. Conversations and case data are private by default. You control visibility with the lock icon for conversations and the Access tab for cases. Code execution runs in your browser via Pyodide — data never leaves your session.
Can I use Purna without loading a patient case?
Yes. In Agent Mode you can search genomic databases, research genes and diseases, build gene panels, and generate visualizations — all without patient data.
What genomic databases does Purna search?
Purna searches databases across five categories: Research Papers (PubMed, Nature, Cell, NEJM, Lancet, arXiv, bioRxiv), Clinical Evidence (ClinVar, ClinicalTrials, OMIM, ClinGen, Orphanet), Variant Databases (gnomAD, GWAS Catalog, dbSNP, VarSome, Franklin), Guidelines (ACMG, NCCN, UpToDate, CDC, WHO), and Drugs & Therapies (PharmGKB, DrugBank, FDA, ClinicalTrials). Toggle sources on/off with the Sources dropdown.
How do gene panels work?
Gene panels are curated gene lists for targeted variant filtering. Create custom panels or search existing ones, then apply them from chat (“Find variants in my cardiomyopathy panel”) or from the Virtual Panels filter in the Variants tab.
Can I share cases with my team?
Yes. In any case, go to the Access tab to add team members with Admin or General roles. Manage broader team settings from Settings > Team.
What is the variant detail panel?
Click any variant in the variant browser to open a detail panel with seven tabs: Assessment (in silico predictions), Details (genomic location, HGVS, population frequencies), Research (ClinVar, PubMed, gene info), ACMG (classification criteria), gnomAD (population data browser), AI Analysis (AI-powered interpretation), and Notes (collaborative annotations).
How does ACMG classification work?
The ACMG tab in the variant detail panel shows the 28 standard ACMG/AMP criteria. Use AI Enhance to have Purna automatically evaluate criteria using ClinVar, PubMed, gnomAD, and functional data. Review suggestions, accept or reject individual criteria, and manually reclassify with Reclassify. All changes are tracked in the reclassification history.
What clinical reports can Purna generate?
Five report types: variant interpretation, clinical findings, case summary, ACMG evidence, and variant comparison. Ask from chat (e.g., “Generate a case summary report”) and Purna creates the document in the artifacts panel for download.
What are findings?
Findings are variants marked as clinically significant during analysis. Categorize each as Primary (explains phenotype), Causal (confirmed cause), Secondary (related but not primary), or Incidental (unrelated but actionable). View and export findings from the Findings tab.
What are HPO phenotypes and how do I use them?
HPO (Human Phenotype Ontology) phenotypes describe a patient’s clinical features. Add them via AI Assist (Purna suggests terms from clinical descriptions) or Search (manual lookup). Phenotypes become powerful filters in the Variants tab as Primary or Secondary Phenotypes.
What is in the case Settings tab?
The Settings tab lets you rename the case, edit case details, upload and manage files, view the settings activity audit log, and archive the case from the Danger Zone section.
What chart types are available?
Pie/donut, bar, horizontal bar, line, and area charts. Describe the visualization you want in natural language and Purna generates it. After generation, ask to adjust colors, labels, or chart type.
How does code execution work?
Purna runs Python in your browser through Pyodide. Available libraries: pandas, matplotlib, and seaborn. Your data stays in your session and never leaves your machine.
Why does a case take a moment to load?
Cases that haven’t been accessed recently may be archived for performance. When you open one, Purna retrieves and loads it. This is typically brief.
Can I export my analysis results?
Yes. Export findings from the Findings tab, generate clinical reports from chat, create spreadsheet artifacts (CSV), and save Python scripts. Reports appear as downloadable artifacts in the side panel.
What plans are available?
Three tiers: Individual (50/month,50queries/day),∗∗Pro∗∗(125/month, API access, background analysis), and Team ($999/month, unlimited queries, SSO, custom branding, unlimited members). Manage from Settings > Plan.
How do I customize how Purna responds?
Go to Settings > General > Personalization and configure Base Style (Detailed/Concise), Tone (Professional/Casual), Characteristics (Balanced/Cautious/Confident), and Output Format (Structured/Conversational).
What case types does Purna support?
Currently Single Inherited Disease. Family Inherited Disease and Somatic Case types are coming soon.
How do citations work?
Responses include inline citation badges (ClinVar, NCBI, PubMed, etc.) that link directly to source records. Click any badge to verify the information.
How do I invite team members?
Go to Settings > Team and click Invite to generate an invitation link. Share the link with your colleagues. They can create an account (if needed) and join your organization through the link.