메뉴
Annotate cell types from normalized scRNA-seq data using marker scoring, CellTypist, PopV-style reference mapping, lightweight KNNPredict-style mapping, SingleR, or scmap through shared Python/R backends.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/mdbabumiamssm/LLMs-Universal-Life-Science-and-Clinical-Skills- --skill sc-cell-annotation명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
Operate MedSAM2 for promptable segmentation of 3D medical images and medical videos, including CT lesion propagation, MRI volumes, RECIST-guided prompts, efficient CPU-oriented variants, training, and 3D Slicer integration. Use when generating or validating volumetric masks from sparse prompts or propagating masks through image slices or video frames.
Build reproducible healthcare imaging pipelines with Project MONAI for DICOM, NIfTI, pathology, and multidimensional imaging tasks including preprocessing, augmentation, training, sliding-window inference, evaluation, model bundles, labeling, and deployment. Use when implementing medical image classification, segmentation, registration, detection, generative, or foundation-model workflows in PyTorch.
Operate Google TxGemma prediction and chat models for therapeutic property prediction across small molecules, proteins, nucleic acids, diseases, targets, and cell lines. Use when formatting Therapeutics Data Commons tasks, choosing TxGemma model size or variant, running local or Model Garden inference, fine-tuning on private therapeutic data, or evaluating TxGemma in drug-discovery workflows.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | sc-cell-annotation |
| description | Annotate cell types from normalized scRNA-seq data using marker scoring, CellTypist, PopV-style reference mapping, lightweight KNNPredict-style mapping, SingleR, or scmap through shared Python/R backends. |
| version | 0.9.0 |
| author | OmicsClaw |
| license | MIT |
| tags | ["singlecell","annotation","celltypist","popv","knnpredict","singler","scmap"] |
| metadata | {"omicsclaw":{"domain":"singlecell","allowed_extra_flags":["--manual-map","--manual-map-file","--cluster-key","--method","--model","--reference","--marker-file","--celltypist-majority-voting","--no-celltypist-majority-voting","--r-enhanced"],"param_hints":{"manual":"[Truncated]","markers":"[Truncated]","celltypist":"[Truncated]","popv":"[Truncated]","knnpredict":"[Truncated]","singler":"[Truncated]","scmap":"[Truncated]"},"legacy_aliases":["sc-annotate"],"saves_h5ad":true,"requires_preprocessed":true}} |
Implemented methods:
manualmarkerscelltypistpopvknnpredictsinglerscmapThis skill annotates cells or clusters. It does not replace marker discovery or replicate-aware DE.
adata.Xsc-clusteringsc-markers, sc-de, or interpretation/reporting--method--manual-map--manual-map-file--cluster-key--model--reference--marker-file--celltypist-majority-votingSuccessful runs write:
processed.h5adreport.mdresult.jsonfigures/embedding_cell_type.pngfigures/embedding_cluster_vs_cell_type.pngfigures/cluster_to_cell_type_heatmap.pngfigures/cell_type_counts.pngfigures/embedding_annotation_score.png when scores are availablefigures/manifest.jsonfigure_data/manifest.jsontables/annotation_summary.csvtables/cell_type_counts.csvtables/cluster_annotation_matrix.csvreproducibility/commands.shreport.mdfigures/embedding_cell_type.pngfigures/embedding_cluster_vs_cell_type.pngtables/annotation_summary.csvprocessed.h5ad| Scenario | Recommended method | Example |
|---|---|---|
| PBMC / immune cells (human) | markers (default) or celltypist | --method markers |
| Immune cells (any tissue, human) | celltypist | --method celltypist --model Immune_All_Low.pkl |
| Mouse data | celltypist with mouse model, or custom markers | --method celltypist --model Mouse_Isocortex_Hippocampus.pkl |
| Specialized tissue (brain, lung, etc.) | celltypist with tissue model, or --marker-file | --method celltypist --model Human_Lung_Atlas.pkl |
| Have a labeled reference dataset | knnpredict or popv | --method knnpredict --reference ref.h5ad |
| R environment available | singler or scmap | --method singler --reference HPCA |
JSON format (recommended):
{
"T cell": ["CD3D", "CD3E", "CD4"],
"B cell": ["MS4A1", "CD79A", "CD79B"],
"Macrophage": ["CD68", "CD163"]
}
CSV format:
T cell,CD3D;CD3E;CD4
B cell,MS4A1;CD79A;CD79B
Macrophage,CD68;CD163
List all available models:
import celltypist
celltypist.models.models_description()
Common models:
Immune_All_Low.pkl / Immune_All_High.pkl — pan-immune (human)Human_Lung_Atlas.pkl — human lungCells_Intestinal_Tract.pkl — human intestinalHuman_AdultAged_Hippocampus.pkl — human brainMouse_Isocortex_Hippocampus.pkl — mouse brainDeveloping_Mouse_Brain.pkl — developing mouse brainPan_Fetal_Human.pkl — human fetal tissuesDownload a specific model:
celltypist.models.download_models(model="Immune_All_Low.pkl")
Your reference must be an H5AD file with:
.Xcell_type column in .obs (or another label column)Sources for labeled references:
CD3D, MS4A1, EPCAMCd3d, Ms4a1, Epcammarkers method will auto-detect species mismatch and attempt case-insensitive matchingcelltypist with a mouse-specific model or provide a custom --marker-file with mouse gene namesmethod and its corresponding model / reference / cluster_key as the key scientific choices.markers, report both requested and executed methods.singler / scmap use celldex atlases, remember they may still fail in restricted-network or empty-cache environments even when R packages are installed.| Flag | Type | Default | Description | Validation |
|---|---|---|---|---|
--input | str | — | Input .h5ad file | required unless --demo |
--output | str | — | Output directory | required |
--demo | flag | off | Run with bundled PBMC3k data | — |
--method | str | markers | Annotation method: manual, markers, celltypist, popv, knnpredict, singler, scmap | validated against METHOD_REGISTRY |
--model | str | Immune_All_Low | CellTypist model name or file stem | celltypist only |
--reference | str | HPCA | SingleR/scmap atlas key or path to labeled H5AD (popv/knnpredict) | — |
--cluster-key | str | None | Cluster/label column for marker summaries; auto-detected when omitted | — |
--manual-map | str | None | Inline cluster-to-label mapping, e.g. 0=T cell;1,2=Myeloid | manual only |
--manual-map-file | str | None | Path to mapping file (json/csv/tsv/txt) | manual only |
--marker-file | str | None | Path to custom marker gene file (JSON or CSV) | markers only |
--celltypist-majority-voting | flag | off | Enable CellTypist neighborhood majority-voting smoothing | celltypist only |
--species | str | Human | SCSA species filter (Human/Mouse) | scsa only |
--tissue | str | All | SCSA tissue filter (e.g. Blood, Brain, All) | scsa only |
--scsa-foldchange | float | 1.5 | SCSA DE fold-change threshold | scsa only |
--scsa-pvalue | float | 0.05 | SCSA DE p-value threshold | scsa only |
--r-enhanced | flag | off | Also render R Enhanced ggplot2 figures | — |
Activated by --r-enhanced. Files written to figures/r_enhanced/.
| Renderer | Output file | figure_data CSV | Plot description | Required R packages |
|---|---|---|---|---|
plot_embedding_discrete | r_embedding_discrete.png | annotation_embedding_points.csv | UMAP/embedding colored by annotated cell types | ggplot2 |
plot_embedding_feature | r_embedding_feature.png | annotation_embedding_points.csv | UMAP/embedding colored by annotation confidence score | ggplot2 |
plot_cell_barplot | r_cell_barplot.png | cell_type_counts.csv | Bar chart of cell type counts | ggplot2 |
plot_cell_proportion | r_cell_proportion.png | cell_type_counts.csv | Stacked proportion bar chart per sample or cluster | ggplot2 |
plot_cell_sankey | r_cell_sankey.png | cluster_annotation_matrix.csv | Sankey/alluvial diagram from cluster to cell type | ggplot2, ggalluvial |
Upstream: sc-clustering Downstream: sc-markers, sc-de, sc-cell-communication, sc-differential-abundance