Cis-regulatory elements are small fragments in the genome, ranging from five to twenty nucleotides in length, that do not encode a gene but control when the cell does or does not transcribe a gene. You can think of them as a gene’s ‘dimmers’. When these elements are known, you can influence when, where, and how often that gene is expressed. If you don’t know them, you can only influence the gene itself, for example by selecting for a functional or non-functional version of the gene.

The problem is that these regulatory elements are hidden within the genome. Many of the cis-regulatory elements currently known in plants have often been identified on a gene-by-gene basis and are relatively young from an evolutionary perspective. Systematic research, involving the analysis of entire genomes, has not yielded much so far.

Yet there were indications that plant genomes, just like those in animals, also contain evolutionarily ancient cis-regulatory elements. There are thousands of genes in plants whose activity is highly conserved, but for which no conserved cis-regulatory elements are known. ‘That is illogical; the only logical explanation was that they were hidden’, says Zachary Lippman of Cold Spring Harbor Laboratory in the U.S. and co-author of the Science study.

Immediate surroundings

To find these hidden elements, the researchers took an unconventional approach. ‘Instead of comparing entire genomes, we focused on genes and their immediate surroundings’, says Lippman. This narrowed the search area by a factor of about 100. The algorithm the group developed for this purpose scanned all the genes of the 284 plants included in their analysis. These were mostly seed plants (274 species); the remaining 10 were ferns and mosses, which are evolutionarily more distant from seed plants. The search resulted in the identification of 2.3 million new, yet evolutionarily ancient (dating back up to 150 million years) cis-regulatory elements.

‘The only logical explanation was that they were hiding.’

Zachary Lippman, Cold Spring Harbor Laboratory

It also became clear why these elements are difficult to find. ‘This study shows that such regulatory elements jump around in different ways. Sometimes the element is located in front of a gene, sometimes behind it, and sometimes another gene is situated in between’, says Klaas Vandepoele, professor of computational regulomics at the VIB-UGent Center for Plant Systems Biology and not involved in the study. ‘So that’s quite interesting, and it also demonstrates that this regulatory code is actually quite flexible, but it always works.’

Piano keys

For now, all that information about the newly identified elements is stored in a separate database. There is not yet a link to databases containing all other information about plant genomes, such as gene functions. “It would certainly make sense to link all that information,” says Vandepoele.

With the identified elements, researchers can delve deeper into the question of how gene activity is regulated. Lippman compares cis-regulatory elements to the keys of a piano. “But which combination of keys is needed to play a particular song on the piano—that remains unknown.” The traditional way to figure that out is by introducing mutations and edits into those elements and studying what happens to gene activity and the phenotype.

But another possibility is to use AI for this. That’s what Vandepoele is working on. You can ask recent AI models which specific segment of the DNA sequence such a model used to predict a gene’s activity. Combine that with this new database of cis-regulatory elements, and ‘you’ve got a very powerful predictive engine’, says Vandepoele. ‘Those AI models can then say, “If you remove this element, your gene will no longer be expressed in, for example, a specific organ.”’

‘Then you’ll have a very powerful predictive engine.’

Klaas Vandepoele, VIB-UGent Center voor Plant Systems Biology

Another area researchers can now explore in greater depth is the rules governing the binding of transcription factors to cis-regulatory elements. Geneticist Dolf Weijers, a professor at Wageningen University & Research and also not involved in the study, has been investigating this for quite some time using a specific group of transcription factors known as auxin response factors. ‘We know the binding mechanism for these very well’, says Weijers. What’s interesting about auxin response factors is that they are ‘ultra-conserved.’ There is no variation in the structure of these proteins from different plants at the sites that bind directly to the DNA. The assumption is that this is also the case on the DNA side. With this new library of elements, they can begin testing whether the rules they have discovered are actually correct.

More targeted breeding

‘It’s actually quite easy for breeding companies to apply that data to the genomes of their own varieties’, says Vandepoele. This allows breeders to check whether their markers for certain traits overlap with a cis-regulatory element. It also makes it easier to identify the regulatory elements of a specific gene. ‘You shouldn’t underestimate that impact’, says Weijers.

‘That way, you can take a much more targeted approach.’

Dolf Weijers, WUR

Take, for example, a breeder working with a specific lettuce variety who wants to alter the expression of an important gene. Until now, that was very difficult without specific knowledge of the regulation in the lettuce variety where the breeder wants to change the expression. ‘But what you can do now is go to that database and see, ah okay, in this variety there are elements here, here, here, and here that are completely conserved, all of which are important for regulation’, says Weijers. ‘Then you can work in a much more targeted way.’ Although it still remains to be figured out exactly what those elements do.

Lippman compares the database of cis-regulatory elements to the improved annotation of genes. Back then, too, researchers were able to ask questions that previously couldn’t be answered. ‘It’s a very exciting time’, says Lippman.

Kirk R. Amundson, Anat Hendelman, et al., A deep-time landscape of plant cis-regulatory sequence evolution, Science (2026) https://www.science.org/doi/10.1126/science.adt898