MEIRLOP: improving score-based motif enrichment by incorporating sequence bias covariates

Abstract

Background Motif enrichment analysis (MEA) identifies over-represented transcription factor binding (TF) motifs in the DNA sequence of regulatory regions, enabling researchers to infer which transcription factors can regulate transcriptional response to a stimulus, or identify sequence features found near a target protein in a ChIP-seq experiment. Score-based MEA determines motifs enriched in regions exhibiting extreme differences in regulatory activity, but existing methods do not control for biases in GC content or dinucleotide composition. This lack of control for sequence bias, such as those often found in CpG islands, can obscure the enrichment of biologically relevant motifs. Results We developed Motif Enrichment In Ranked Lists of Peaks (MEIRLOP), a novel MEA method that determines enrichment of TF binding motifs in a list of scored regulatory regions, while controlling for sequence bias. In this study, we compare MEIRLOP against other MEA methods in identifying binding motifs found enriched in differentially active regulatory regions after interferon-beta stimulus, finding that using logistic regression and covariates improves the ability to call enrichment of ISGF3 binding motifs from differential acetylation ChIP-seq data compared to other methods. Our method achieves similar or better performance compared to other methods when quantifying the enrichment of TF binding motifs from ENCODE TF ChIP-seq datasets. We also demonstrate how MEIRLOP is broadly applicable to the analysis of numerous types of NGS assays and experimental designs. Conclusions Our results demonstrate the importance of controlling for sequence bias when accurately identifying enriched DNA sequence motifs using score-based MEA. MEIRLOP is available for download from https://github.com/npdeloss/meirlop under the MIT license.

Glucocorticoid Receptor-Regulated Enhancers Play a Central Role in the Gene Regulatory Networks Underlying Drug Addiction

“…To gain insights into the TFs that may mediate changes in gene expression networks in response to a history of substance abuse, we identified TF motifs enriched in TSRs regulated by oxycodone or cocaine exposure in each brain region. We used MEIRLOP ( Brigidi et al, 2019 ; Delos Santos et al, 2020 ), a DNA motif analysis approach that associates motifs with the magnitude of regulation at TSRs across conditions based on logistic regression. This analysis identified a strong and consistent association between the glucocorticoid response element (GRE) and TSRs down-regulated in brain tissue from rats with addiction-like phenotypes versus controls ( Figure 3B and Supplementary Figure 4A ).…”

“…To analyze motif enrichment associated with changes in transcription levels, we analyzed regulated TSRs with MEIRLOP ( Delos Santos et al, 2020 ). Sequences were scored based on their shrunken log2 fold change between treatment conditions (e.g., naive vs. cocaine or oxycodone exposed) and analyzed with MEIRLOP using HOMER’s known transcription factor motif library.…”

Benchmarking tools for transcription factor prioritization

“…In this benchmark study, we examined the performance of nine TF prioritization tools in combination with the two most common parameters (PWM motif library and the set of background sequences used), resulting in 13 approaches to rank TFs [16][17][18][19][20][21][22][23][24]. The ground truth for this was defined using a collection of published H3K27ac ChIP-seq experiments which included a TF perturbation in their design (OE, KO, etc.).…”

“…To this end, computational tools have been developed to perform TF prioritization based on different assumptions and implementations [16][17][18][19][20][21][22][23][24]. Among these, we could broadly identify two types, depending on their underlying reference: 1) tools leveraging DNA sequence information using position weight matrices (PWMs) to predict TF binding (PWM based tools), and 2) sequence-independent tools using previously identified TF binding sites in the genome (ChIP-seq peak based tools).…”

“…Among these, we could broadly identify two types, depending on their underlying reference: 1) tools leveraging DNA sequence information using position weight matrices (PWMs) to predict TF binding (PWM based tools), and 2) sequence-independent tools using previously identified TF binding sites in the genome (ChIP-seq peak based tools). Independently of their reference, both types of tools are prioritizing TFs based on statistical methods such as Fisher's exact test, rank based enrichment, and LASSO regression, among others [16][17][18][19][20][21][22][23][24].…”