Skip to main content

I-Impute: a coherent strategy to impute singlecell RNA sequencing data

Project description

I-Impute: a coherent strategy to impute singlecell RNA sequencing data

I-Impute is a coherent method to impute scRNA-seq data. I-Impute leverages continuous similarities and dropout probabilities and refines the data iteratively to make the final outputself-consistent. I-Impute exhibits robust imputation ability and follows the coherent principle. It offers perspicacity to uncover the underlying cell subtypes in real scRNA-Seq data.


  • python3
  • numpy>=1.16.1
  • pandas>=0.23.4,<0.24
  • scipy>=1.3.0
  • scikit-learn>=0.21.1
  • tasklogger>=0.4.0

install requirements

pip install -r requirements.txt


Installation with pip

To install with pip, run the following from a terminal:

pip install i-impute

Installation from Github

To clone the repository and install manually, run the following from a terminal:

git clone
cd I-Impute
python install


Quick start

The following code runs MAGIC on simulation data located in the I-Impute repository.

import iimpute
import pandas as pd

# read your reads count or RPKM or TPM data
data = pd.read_csv('simluation-data/sim-counts.csv', index_col=0)

# create I-Impute object
iimpute_operator = iimpute.IImpute(normalize=False)

# impute
imputed_data = iimpute_operator.impute(data)

# store result to a file
imputed_data.to_csv('your file name')

# iterative mode
iimpute_operator = iimpute.IImpute(normalize=False, iteration=True)

# impute
imputed_data = iimpute_operator.impute(data)

# store result to a file
imputed_data.to_csv('your file name')


IImpute(n=20, c_drop=0.5, p_pca=0.4, alpha=0.01, normalize=True, iteration=False, verbose=1)


  • n : int, optional, default: 20

    The nth of nearest neighbors on which to build kernel when calculating affinity matrix.

  • c_drop : float, optional, default: 0.5

    Dropout event cutoff. For entry whose dropout probability is less than c_drop, we consider it as a real observation, its original value will remain. Otherwise, we conduct the imputation with the aid of information from similar cells.

  • p_pca : float, optional, default: 0.4

    Percentage of variance explained by the selected components of PCA. It determines the nmumber of PCs used to calculate the distance between cells.

  • alpha : float, optional, default: 0.01

    L1 penalty for Lasso regression.

  • normalize : boolean, optional, default: True

    By default, I-Impute takes in an unnormalized matrix and performs library size normalization during the denoising step. However, if your data is already normalized or normalization is not desired, you can set normalize=False.

  • iteration : boolean, optional, default: False

    The imputation process only performs once when False (it is equivalent to C-Impute described in our paper). The imputation process will iterate n times to achieve self-constistent imputation matrix.

  • verbose : int or boolean, optional, default: 1

    If True or > 0, print status messages

Cite us



If you have any questions or require assistance using I-Impute, please contact us with

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

i-impute-1.0.1.tar.gz (7.1 kB view hashes)

Uploaded source

Built Distribution

i_impute-1.0.1-py3-none-any.whl (8.2 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page