Human DNA “Codebook” maps 177 transcription factors—and shows methylation can change the rules

Two studies published as part of the same international research effort sharpen the picture of how human cells regulate genes. The first, led by Timothy Hughes at the University of Toronto, created an expanded “Codebook” of transcription-factor DNA-binding preferences. The companion work from Bart Deplancke’s team introduced a way to test how DNA methylation changes those preferences.

The central finding is that gene regulation depends on more than the underlying DNA sequence. Methylation can alter how a transcription factor recognizes a sequence—sometimes weakening binding, sometimes changing the preferred sequence, and sometimes increasing affinity.

What the Codebook measured

Transcription factors are proteins that help control gene expression by recognizing short DNA sequences called motifs. Many human transcription factors still had little or no experimentally established binding information.

The Codebook consortium combined five experimental platforms with computational analysis and carried out more than 4,800 experiments on 332 putative or poorly characterized human transcription factors. It identified DNA-binding motifs for 177 of them, or 53%, adding around 130 distinct motifs to the known vocabulary of human gene regulation.

The newly measured motifs were strongly enriched at transcription-factor binding sites observed in cells. That allowed the researchers to identify tens of thousands of previously unrecognized, conserved and direct genomic binding sites, with many concentrated in promoter regions. The data also improved the ability to connect transcription-factor binding with gene-expression patterns.

The resulting resource expands motif coverage to about 1,421 of an estimated 1,600 human transcription factors. It is not a complete explanation of gene regulation, but it substantially reduces a major gap in the reference map: knowing which proteins can recognize which DNA sequences.

What meSMiLE-seq adds

The companion study focused on a question that ordinary motif maps cannot answer: does a transcription factor bind differently when a DNA base is chemically modified?

The researchers introduced meSMiLE-seq, or methylation-sensitive selective microfluidics-based ligand enrichment followed by sequencing. The microfluidic assay compares a transcription factor’s binding to matched methylated and unmethylated DNA, making it possible to model both sequence preference and methylation sensitivity.

The dataset record reports tests of 114 human transcription factors across 23 experiments, producing binding models for 48. Among those 48, 13 showed aversion to methylated DNA, while 11 showed either a preference for methylated DNA or an alternative binding site that included the modification. Binding for the remaining factors was unaffected under the assay conditions.

This breakdown matters because some summaries give the methylation-preference or altered-binding group as 14. The cited dataset abstract reports 11, so the more cautious interpretation is that methylation-dependent effects were observed in a subset of factors, with 13 showing reduced binding and 11 showing increased or altered preference in that record.

Methylation is not a universal off switch

The results challenge a simplified view of DNA methylation as an automatic repression signal. Methylation can interfere with a transcription factor’s recognition of a motif, but it can also create a preferred chemical context or redirect which sequence a factor binds. The effect depends on the transcription factor and the DNA sequence being tested.

That helps explain how cells with the same genome can maintain different regulatory programs. Cell types differ in the transcription factors they produce and in the locations where DNA methylation is present. As a result, the same sequence may be ignored in one cellular context, recognized by a different factor in another, or bound with a different strength when methylation changes the local DNA signal.

The studies therefore connect two layers of gene regulation: the sequence code written in the DNA letters and the epigenetic state that changes how those letters are read.

Why the map could matter for disease research

The expanded motif catalogue can improve analysis of noncoding genetic variants—changes that occur outside protein-coding regions but may affect gene regulation. A variant might destroy an existing transcription-factor motif or create a new one; methylation could then determine whether the altered site is actually recognized in a particular cell type.

That makes the resource potentially useful for studying regulatory changes involved in development, aging and cancer. Its immediate role, however, is as a foundational research reference rather than a clinical diagnostic test. The binding models provide a stronger starting point for experiments and interpretation, but biological effects still depend on cellular context and must be validated in relevant systems.

The broader significance

The Codebook project represents a large collaboration spanning more than two dozen institutions across North America, Europe and Russia. Its value is not only the number of factors measured, but also the combination of complementary assays, computational analysis and cellular validation.

Together, the studies show why a more realistic model of gene regulation is needed. Human cells do not read DNA through a fixed dictionary in which every motif has one immutable meaning. They interpret sequence alongside chemical modifications, available transcription factors and the surrounding cellular environment. The Codebook begins to catalogue that sequence vocabulary, while meSMiLE-seq shows how methylation can change the way the vocabulary is read.