Infer gene regulatory networks from time-series single-cell CRISPR data.
Project description
RENGE
RENGE infers gene regulatory networks (GRNs) from time-series single-cell CRISPR screening data.
Install
pip install renge
Usage
Network inference
reg = Renge()
A = reg.estimate_hyperparams_and_fit(X, E)
A_qval = reg.calc_qval()
input
E : C x G pandas DataFrame of expression data. The expression data should be normalized and log-transformed.
X : C x (G+1) pandas DataFrame. The rows of X[:, :G] are one-hot vectors indicating the knocked out gene in each cell. The last column of X indicates the sampling time of each cell.
Here,
C : The number of cells.
G : The number of genes included in the GRN.
output
A : G x G pandas DataFrame. (i, j) element of A is a regulatory coefficient from gene j to gene i.
A_qval : G x G pandas DataFrame. (i, j) element of A_qval is a q-value for (i, j) element of A.
Prediction of expression changes after gene knockout
reg = Renge()
reg.estimate_hyperparams_and_fit(X_train, E_train) # train the model
E_pred = reg.predict(X_pred)
input
E_train : C x G pandas DataFrame of expression data. The expression data should be normalized and log-transformed.
X_train : C x (G+1) pandas DataFrame. The rows of X_train[:, :G] are one-hot vectors indicating the knocked out gene in each cell. The last column of X_train indicates the sampling time of each cell.
X_pred : T x (G+1) pandas DataFrame. The rows of X[:, :G] are real-valued vectors indicating expression change of target gene of perturbation. For knockout or knockdown, values should be negative. The last column of X_pred indicates the time at which expressions are predicted
Here,
T : The number of timepoints where expressions are predicted.
output
E_pred : T x G pandas DataFrame of predicted expression.
Reference
TBD
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for renge-0.0.2.dev3-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | b00fd127b917027cdb5c93b58a4c789fbe94cf9376165b2aa87faeebd4b0af22 |
|
MD5 | 3ba6547919333d1fc611322d488b1d73 |
|
BLAKE2b-256 | 3a2051392f6bd8fe9bbfcc13d06360e6f97c5bd8e6fd14388bdea474b67af304 |