Multiomics Agent
Omi's specialised co-pilot for multiomics work
I integrate your single-cell modalities — scRNA, scATAC, CITE-seq, methylation — into one coherent picture. Peak-to-gene links, motif enrichment, joint embeddings: I handle the messy plumbing so you can focus on the biology.
What I can do for you
I load multimodal data (10x Multiome, MuData, ArchR, Signac objects) and build joint embeddings using MOFA+, WNN, totalVI, or GLUE — whichever fits your design best, and I'll explain why.
I compute peak-to-gene links across your cell types, run motif enrichment with chromVAR or Homer, and surface candidate regulators (TFs whose motif accessibility tracks with their target gene expression).
I integrate CITE-seq protein with RNA, regress out modality-specific noise, and call cell types using both layers — so a CD4 that's also got high CD25 protein gets called Treg, not just CD4 naive.
I run cross-modality differential analyses: which peaks are differentially accessible between two clusters, and which of those peaks have a linked gene that's also DE? That's where the real regulatory biology lives.
Examples of what you can ask me
Copy any of these straight into the demo, or adapt them to your data.
- 1"Integrate my scRNA and scATAC and build a joint UMAP."
- 2"Find peak-to-gene links in my CD8 effector cluster."
- 3"Run motif enrichment on differentially accessible peaks between Treg and Tconv."
- 4"Which transcription factors are driving my macrophage polarization?"
- 5"Build a WNN embedding from my CITE-seq antibody panel + RNA."
- 6"Compare gene activity scores from ATAC vs RNA expression for IRF8."
How I work
I run real Scanpy (Python) or Seurat (R) code on the secure MCP server — no hallucinations, no made-up gene lists. Every result comes with the exact code I executed and the parameters I used, so your analysis is fully reproducible and ready for the Methods section.
Best for
Anyone running 10x Multiome, CITE-seq, or combining separately-assayed modalities — especially developmental biologists, immunologists, and gene-regulation folks chasing TF circuits.
References
- MOFA+ (Argelaguet et al., 2020) – Genome Biology
- Seurat WNN (Hao et al., 2021) – Cell
- totalVI (Gayoso et al., 2021) – Nature Methods
- GLUE (Cao & Gao, 2022) – Nature Biotechnology
- Signac (Stuart et al., 2021) – Nature Methods
- ArchR (Granja et al., 2021) – Nature Genetics
- chromVAR (Schep et al., 2017) – Nature Methods
Try Multiomics now
Jump into the demo with a starter prompt already loaded. Upload your data, or play with our example dataset first.