Transform raw sequencing counts into biological insights within minutes using deep learning, automated pipelines and scalable cell annotation.
from cellscribe import run_cellscribe
# Load your single-cell dataset
results = run_cellscribe(
file_path="pbmc3k.h5ad",
use_scvi=True,
annotate=True
)
# Access cell type annotations
cell_types = results.annotations
umap = results.embeddings
# Generate publication-ready figures
results.plot_umap(color="cell_type")
results.plot_marker_genes(n_genes=10)A complete toolkit designed for computational biologists and bioinformatics researchers.
Seamlessly import .h5ad, .loom, 10x Genomics, and Seurat objects with automatic format detection.
Intelligent quality control with adaptive thresholds for mitochondrial content, doublet detection, and gene counts.
scVI-powered variational autoencoders for batch correction, denoising, and latent representation learning.
Automated cell annotation using CellTypist with 42+ immune cell types and tissue-specific reference atlases.
Statistically robust DE analysis with multiple testing correction, volcano plots, and pathway enrichment.
Generate high-resolution UMAPs, heatmaps, violin plots, and dot plots ready for Nature, Cell, and Science journals.
From raw counts to biological insights in a single, reproducible workflow.
Import .h5ad, .loom, or 10x data
Filter low-quality cells & genes
Log1p & size factor correction
Highly variable gene detection
Deep latent space modeling
Non-linear dimensionality reduction
Graph-based cell clustering
AI-powered cell annotation
Differential expression testing
Interactive HTML output
Validated on real-world datasets with ground-truth annotations.
| Dataset | Cells | Runtime | Accuracy | Status |
|---|---|---|---|---|
|
PBMC3K
|
3,000 | 2 min |
88%
|
Validated |
|
PBMC68K
|
68,000 | 4.7 min |
84%
|
Validated |
|
Human Lung Atlas
|
120,000 | 7 min |
86%
|
Validated |
Leveraging the best of the single-cell ecosystem.
Upload datasets, discover cellular heterogeneity and accelerate biological discovery.