In RNA-Seq data analysis, once a set of differentially expressed genes has been identified, the next important step is to interpret those changes biologically.
In Subio Platform’s
RNA-Seq Data Analysis Tutorial,
“Chapter 7: Gene Annotation and Enrichment Analysis — From Statistical Results to Biological Interpretation”
introduces enrichment analysis using transcription factor target gene sets provided by GSEA.
In addition,
“Chapter 8: Genes Regulated in a Genomic Position-Specific Manner and Motif Sequences”
shows how to search for genes containing a particular motif sequence within a specified distance from the transcription start site (TSS).
However, both approaches are based on sequence information or predicted gene lists. The presence of a motif sequence does not necessarily mean that the transcription factor (TF) actually binds there in the relevant cell type or experimental condition. Similarly, a predicted target gene list does not directly demonstrate actual binding or transcriptional regulation.
In this article, we use ChIP-Atlas, a database that integrates public ChIP-seq data, to show how RNA-Seq expression patterns can be interpreted in relation to transcription factor binding information.
What Is ChIP-Atlas?
ChIP-Atlas is a data-mining suite that integrates publicly available experimental data, including ChIP-seq, ATAC-seq, and Bisulfite-seq, and allows users to explore epigenomic information. The ChIP-Atlas website explains that it integrates hundreds of thousands of ChIP-seq, ATAC-seq, and Bisulfite-seq experiments.
ChIP-Atlas provides several tools, including Peak Browser, Target Genes, Enrichment Analysis, Colocalization, and Dataset Search. In this video, we mainly use the following two tools.
| Tool | Main purpose | How it is used here |
|---|---|---|
| Peak Browser | View genomic peaks for a particular TF, histone modification, or other feature | Download ESRRA ChIP-seq peaks as a BED file and import them into Subio Platform |
| Enrichment Analysis | Identify TFs or other features whose peaks are enriched near an input set of genes or genomic regions | Explore candidate TFs whose binding peaks are enriched near genes with increased expression |
How Should ChIP-Atlas Results Be Interpreted?
ChIP-Atlas allows us to use transcription factor binding peaks observed in actual ChIP-seq experiments as clues for identifying regulatory factors that may lie upstream of gene expression changes.
However, several points must be considered when interpreting ChIP-Atlas results. Transcription factor binding can vary greatly depending on cell type, stimulation conditions, differentiation state, time point, and chromatin state. Therefore, a peak in a particular region is evidence that binding was observed under a specific experimental condition in the past. It does not mean that the TF always binds to that region.
In addition, because ChIP-Atlas integrates public datasets, the antibodies, sequencing depth, peak detection sensitivity, and overall data quality are not uniform across experiments. When a peak is found in ChIP-Atlas, it can provide a useful clue that the TF may be involved. On the other hand, even when no peak is found, the relevant cell type or condition may not be represented in the database, or the peak may not have been detected because of limitations in experimental sensitivity or data quality.
In other words, ChIP-Atlas is highly useful for finding examples in which binding has been observed, but it cannot be used to prove that binding does not occur.
Obtaining ESRRA Binding Peaks with Peak Browser
In the first half of the video, we use the ChIP-Atlas Peak Browser to download ESRRA binding peaks detected in breast-derived cells as a BED file.
The downloaded BED file can be imported into Subio Platform as a Region List. By importing the peaks as a Region List, their genomic positions can be examined together with RNA-Seq data and gene annotations.
Here, we narrowed the ESRRA peaks to those detected in experiments using MDA-MB-231 cells. We then visualized their positions on the hg38 genome and used the Genome Location Filter to extract genes with ESRRA peaks near their TSSs.
As an example, we extracted genes with peaks within a region centered 500 bp upstream of the TSS and extended by 500 bp on either side. In this way, Subio Platform allows users to specify positions relative to the TSS and search for genes with peaks in a particular genomic region.
Examining the Expression Patterns of Genes with ESRRA Peaks Near Their TSSs
The extracted gene list can be used directly to examine expression patterns in the RNA-Seq data. When expression data are visualized in the Genome Browser, genes located close together on the genome may sometimes show similar expression patterns.
This may provide a clue that not only ESRRA-dependent regulation, but also epigenetic regulation associated with chromatin state or genomic position, could be involved. Of course, this observation alone does not demonstrate a causal relationship. However, viewing genomic positions, transcription factor binding peaks, and expression patterns in the same window makes it easier to identify gene groups that may deserve closer examination.
Searching for Candidate Upstream TFs with Enrichment Analysis
In the second half of the video, we perform the analysis in the opposite direction. Starting from a group of genes with a particular expression pattern, we search for candidate transcription factors whose binding peaks are enriched near those genes.
Here, we extracted genes whose expression increased consistently after treatment with three different ESRRA siRNAs and entered their Gene Symbols into the ChIP-Atlas Enrichment Analysis tool. Enrichment Analysis examines which TFs or epigenomic features have peaks that overlap frequently with an input set of genes or genomic regions.
This is not an analysis that simply checks whether a particular TF peak exists near one individual gene. Instead, it examines which TF binding peaks are frequently found near a group of genes showing a particular expression pattern and is used to explore candidate TFs that may be involved upstream.
Examining Enrichment Analysis Results in Subio Platform
The results of ChIP-Atlas Enrichment Analysis can be downloaded as a TSV file. In the video, we opened the result in Excel and narrowed it to experiments using MDA-MB-231 cells. We then extracted the Gene Symbol column, saved it as a tab-delimited file, and imported it into Subio Platform as a gene list.
ESRRA was included in the list of candidate upstream TFs for the genes whose expression increased consistently. Although ESRRA expression itself decreased after ESRRA siRNA treatment, many ESRRA binding peaks were found near the genes whose expression increased consistently. This result provides a clue for considering a possible relationship between the decrease in ESRRA and the increased expression of these genes.
Why Combine ChIP-Atlas with RNA-Seq?
RNA-Seq analysis can reveal which genes show increased or decreased expression and which pathways or GO terms may be involved. However, another type of information is needed to investigate which TFs may be acting upstream of those expression changes.
Using ChIP-Atlas, public ChIP-seq data can be used to examine which TF binding peaks are enriched near differentially expressed genes. This makes it possible to interpret expression patterns identified by RNA-Seq in relation to TF binding and chromatin state information.
Although it is not covered in this video, the Colocalization tool can be used to identify other TFs that tend to bind to the same genomic regions as ESRRA. When such a TF is identified, it becomes possible to consider whether the transcriptional program associated with that TF may have a relatively stronger influence after ESRRA knockdown. The Target Genes tool can then be used to investigate candidate target genes of the TF identified through Colocalization. However, Colocalization analysis alone does not demonstrate that the TF’s binding actually increases after ESRRA knockdown.
By combining ChIP-Atlas with Subio Platform, researchers can develop hypotheses about the upstream mechanisms underlying expression changes observed by RNA-Seq and progressively narrow down candidate TFs and target genes. Together, these tools provide a powerful approach for generating hypotheses about the regulatory mechanisms behind gene expression changes.