Interpretable machine learning models for single-cell ChIP-seq imputation

Abstract

Single-cell ChIP-seq analysis is challenging due to data sparsity. We present SIMPA (https://github.com/salbrec/SIMPA), a single-cell ChIP-seq data imputation method leveraging predictive information within bulk ENCODE data to impute missing protein- DNA interacting regions of target histone marks or transcription factors. Machine learning models trained for each single cell, each target, and each genomic region enable drastic improvement in cell types clustering and genes identification.The discovery of protein-DNA interactions of histone marks and transcription factors is of high importance in biomedical studies because of their impact on the regulation of core cellular processes such as chromatin structure organization and gene expression. These interactions are measured by chromatin immunoprecipitation followed by high-throughput sequencing (ChIP-seq). Public data from the ENCODE portal, which provides a large collection of experimental bulk ChIP-seq data, has been used for comprehensive investigations revealing insights into epigenomic processes impacting chromatin 3Dstructure, open chromatin state, and gene expression to name just a few (ENCODE project consortium, 2012). Recently developed protocols for single-cell ChIP-seq (scChIP-seq) are

Deep Learning Applications in Single-Cell Omics Data Analysis

“…ChIP-seq (chromatin immunoprecipitation followed by sequencing) is a widely used technique for mapping transcription factors, histone changes, and other protein-DNA interactions for genome-wide mapping (Furey 2012). ChIP-seq data is suffering from sparsity hence Albretch et al introduced SIMPA (Albrecht, Andreani et al 2021), a Single-cell ChIP-seq imputation algorithm that was tested on a scChIP-seq dataset of the H3K4me3 and H3K27me3 histone marks in B-cells and T-cells. Unlike most SC imputation approaches, SIMPA integrates the sparse input of one single cell with a series of 2,251 ENCODE ChIP-seq experiments to extract predictive information from bulk ChIP-seq data.…”

“…ChIP-seq data is very sparse, therefore generally requiring imputation for more accurate analysis. Albretch et al introduced Single-cell ChIP-seq iMPutAtion (SIMPA) (Albrecht, Andreani et al 2021), an imputation algorithm that was tested on a single-cell ChIP-seq (scChIP-seq) dataset of the H3K4me3 and H3K27me3 histone marks in B-cells and T-cells. Unlike most SC imputation approaches, SIMPA integrates the sparse input of one single cell with a series of 2,251 ENCODE ChIP-seq experiments to extract predictive information from bulk ChIP-seq data.…”