Description

This package provides functions to analyse missing value mechanisms in the context of bottom-up MS-based quantitative proteomics.

It allows estimating a mixture model of missing completely-at-random (MCAR) values and missing not-at-random (MNAR) values.

It also contains functions allowing the imputation of missing values under hypotheses of MCAR and/or MNAR values.

The main functions of this package are the estim.mix (estimation of a model of MCAR and MNAR (left-censored) values), impute.mi (multiple imputation) and impute.mix (imputation based on a decision rule). It provides also several imputation algorithms for MS-based data. They can be used to impute matrices containing peptide intensities (as Maxquant outputs for instance).

Missing values has to be indicated with NA and a log-2 transformation of the intensities has to be applied before using these functions.

More explanations and details on the functions of this package are available in Giai Gianetto Q. et al. (2020) (doi: doi: 10.1101/2020.05.29.122770).

Author(s)

Maintainer: Quentin Giai Gianetto <quentin2g@yahoo.fr>

References

Giai Gianetto, Q., Wieczorek S., Couté Y., Burger, T. (2020). A peptide-level multiple imputation strategy accounting for the different natures of missing values in proteomics data. bioRxiv 2020.05.29.122770; doi: doi: 10.1101/2020.05.29.122770