celltypist-dev 0.1.10

A tool for semi-automatic cell type annotation

Celltypist

Automated cell type annotation for scRNA-seq datasets

Install

pip install celltypist-dev

Usage

Python package

Sample data and default model

import celltypist

sample_input = celltypist.samples.get_sample_csv()
result = celltypist.annotate(sample_input)
result.predicted_labels_as_df().to_csv("labels.csv")

Using your own data

import celltypist

input_data = "/path/to/cell_by_gene_matrix.csv"
result = celltypist.annotate(input_data)
result.write_excel("annotation_result.xlsx")

Using included models

Included models are downloaded automatically inside a inscript when the defualt model tries to be loaded or by using celltypist --update-models from the command line.

List models included in with the package

import celltypist

for m in celltypist.models.get_all_models():
    print(f"model = '{m}'")

Use the default model by name

import celltypist

input_data = "/path/to/cell_by_gene_matrix.csv"
default_model = celltypist.models.get_default_model()
result = celltypist.annotate(input_data, model=default_model)
result.write_excel("annotation_result.xlsx")

Use a custom model name from the model list

import celltypist

input_data = "/path/to/cell_by_gene_matrix.csv"
result = celltypist.annotate(input_data, model="Immune_v5_allData_lowest_all.pkl")
result.write_excel("annotation_result.xlsx")

Use custom models

import celltypist

indata = "/path/to/cell_by_gene_matrix.csv"
custom_model = "/path/to/custom_model.pkl"

result = celltypist.annotate(input_data, model=custom_model)
result.write_excel("annotation_result.xlsx")

Full example

import celltypist

indata = "/path/to/cell_by_gene_matrix.csv"
custom_model = "/path/to/custom_model.pkl"

result = celltypist.annotate(input_data, model=custom_model)
result.write_excel("annotation_result.xlsx")

print(result.summary_as_df())
result.predicted_labels_as_df().to_csv("labels.csv")
result.probability_matrix_as_df().to_csv("prob_matrix.csv")

Command Line

Basic usage

celltypist --indata=/path/to/dataset.csv

Advance usage

celltypist
    --indata /path/to/dataset.csv \ # input dataset
    --model /path/to/model.pkl    \ # path to model
    --outpref Dataset1_           \ # add a prefix to the output files
    --outdir /path/to/output      \ # set an output directory for the files