* = tied for number of citations

** = some mutant alleles have kernel phenotypes.

If you want to become one of the famous mutant corn genes, it helps if you have an effect that is visible in corn kernels instead of only from fully grown plants.

And here is why:

• A geneticist could determine that the version of c1 that creates yellow kernels is recessive to the version that creates purple kernels just from looking at the ear of corn on left.
• Furthermore, they could tell you that both the male parent (the plant that provided the pollen) and the female parent (the plant on which the ear of corn grew) were both heteryzygous for the c1 genes (they each had one dominant version of the genes and one recessive version), and therefore the corn kernels the parent plants were grown from were both purple.
• They would know with certainty that all of the yellow kernels contain two recessive versions of the c1 gene.
• While they couldn’t predict with absolute certainty whether a specific purple corn kernel on that ear carried two dominant versions of the c1 gene or one dominant and one recessive version, they would know there was a 1/3 chance that kernel has two dominant copies, and a 2/3 chance it had one dominant and one recessive copy.
• That geneticist could make all sorts of predictions about what ears would look like in future generations depending on what colors of corn kernels were planted and which plants were mated with each other.

All this from a single picture of an ear of corn. For a phenotype seen in corn plants but not in kernels (like Knotted1), a geneticist would have to plant a row or more of corn seeds from an ear and examine the growing plants to get the same quantity of information.

And that is why mutations with kernel phenotypes have been so popular over a century of maize genetics research.

But where was I? Oh yeah, gene counts. 32,690 high confidence genes*. Of those, how many have been studied individually?

While I don’t know that anyone knows the precise answer to that question, one indicator is how many maize genes were named before the maize genome was sequenced. People have been naming maize genes since before even the structure of DNA was known, based on the effect mutant version of the gene have on corn plants (for example: waxy1 or yellow stripe1). Later names might be based on the function of the gene (alcohol dehydrogenase1 or superoxide dismutase4), or anything else we know about the gene (wound induced protein1 or male flower specific18). The point being, if someone bothered to name a gene sometime during the last century of maize genetics, it was likely because they were studying it (to a greater or lesser extent). MaizeGDB keeps records of most of the named genes in maize and (excluding chloroplast and mitochondrial genes) I was able to find records of 1181 named genes in maize.

That’s less than 4% of the number of high confidence genes found within the maize genome, and at least a few of the named genes aren’t found within that group (see the first footnote for more details). Why is that number so low?

1. Each of the genes that has been studied in any detail probably represents some grad student’s doctoral thesis. While the tools have gotten better, the expectations for what is involved in characterizing a gene have risen too. I don’t have any statistics on how many maize genetics students earn their PhDs every year (and many of them will have worked on other kinds of projects than characterizing some new mutant gene), but it’s certainly not the thousands that would be required to characterize every gene in the genome in a short period of time.
2. Perhaps more importantly, the first genes to be studied are the ones with the best mutant phenotypes. To be a good mutant to study, breaking a gene should create something obviously different about the plant (it’s purple, or the tassel produces seeds like an ear instead of pollen, or the plants grow along the ground instead of standing upright), but not be so vital that embryos containing broken versions of the gene don’t develop at all. From a project that to knock out every gene in another plant Arabidopsis thaliana we know that many genes can be broken without any obvious effect on the plants that carry broken copies. That doesn’t mean there won’t be still be interesting things wrong with the plants when they’re studied in more detail, but such mutants were less likely to be identified early on. As for genes mutations are usually lethal, they can be studied (a friend in a lab downstairs is working with just such a mutant) but it certainly adds a whole new layer of difficulty to any research project so the genes better be involved in something interesting enough to justify the extra pain and suffering involved.

Now the situation isn’t nearly as grim as it might sound. Nature re-uses related genes over and over again both between and within species, so any time a researcher studies a new gene in detail, that information doesn’t just inform our knowledge of one particular gene in one particular species. Like a candle in a dark room, the information created by the study of a single gene will illuminate, to a greater or lesser extent, nearby genes (genes that have similar sequences to the gene being studied directly.) So even for a gene that’s never been studied in maize, we can make guesses about its function based on any related genes that have benefited from detailed study (either other genes in maize, in other grasses like barley or rice, other plants like arabidopsis or snapdragon, or even in animals or bacteria). While no geneticist worth their pollenating apron wouldn’t need experimental data before being CERTAIN of a gene’s function, knowing something about the functions of related genes is an excellent starting point.

I just finished some “free time” science looking at the classical genes of maize genetics (which displaced the time I normally spend writing for this site), so expect a couple more posts on related topics later this week.

*The good folks at maizesequence.org also produced a set of all the sequences they thought MIGHT be genes which, in addition to the filtered genes, includes ~70,000 more sequences that might or might not be genes. Many of these potential genes are computationally predicted, by programs that look at the underlying characteristics of the DNA sequence itself (how they work is outside my expertise and above my pay grade), but I can personally vouch for the fact that at least some of those “possible” maize genes are the real thing so the true number of genes contained within the maize genome is at least somewhat greater than the 32,690 reported with high confidence. This fact isn’t in any way a criticism of the people involved in sequencing and annotating the maize genome. The vast majority of the high confidence genes (called the filtered gene set) are real, and most of the other 70,000 genes (those included only in the working gene set (which also includes the genes from the filtered gene set)) are probably figments of a computer program’s imagination. Anywhere they chose to draw the line between the two groups was going to put some genes in the wrong category, and they did everything they could to minimize those miscategorizations.

**This doesn’t mean that the genes don’t have important jobs. You can imagine, for example, that genes involved in a plant’s ability to survive disease, water shortages, cold stress or heat stress all won’t create obvious problems for plants grown in the relatively pampered conditions we biologists try to provide for our research subjects when we aren’t actively studying what happens when we stress plants.

Estimates at the time ranged from 40,000 to 150,000 genes. Then the draft genome was published and estimates drop to 30,000-40,000 range. The final genome paper comes out and estimates drop even further. Fast forward to today and the latest annotations show the human genome contains a mere 23,000 genes. I’m sorry. I know we want to feel special, the very pinnacle of evolution.* But at least when it comes to gene count, we’re not. If comparing our own genome to that of Arabidopsis thaliana (the first plant genome to be sequenced, contained 27,000 genes) didn’t drive that home, the second plant genome sequenced, rice, with its 40,577 non-transposon related genes surely made the point.

Yet somehow every time a new plant genome comes out, as the corn genome did this week, it’s big news that it has more genes than our own. Maybe this is partially because journalists have trouble understanding the true impact of this research and it’s an easy fact to latch on to. But I think at least part of it is that we still take it personally that our genomes don’t respect our special-ness as a species.

If it makes you feel any better, think about it this way. Plants have to observe and react to their environments intelligently in order to survive. So does every living thing. But plants don’t contain a single nerve cell. They can’t learn. edit: (with some caveats see the comment section). Every aspect of every reaction to every stimuli, from lack of water, to out competing neighbors, to an excess of boron (yes, there are people who study genetic variation in plant’s ability to survive in high boron soils) must be hard coded into their DNA. It’s one of the big reasons I love studying plant genomics. Everything about a plant is in there, we just have to know enough to understand it. (Though my hard learned advice is not to spend too much time talking about how wonderful it is not to have to worry about cognitive science and neurobiology when dealing with people who work in animal systems.)

*Nevermind that evolution doesn’t actually have a pinnacle just countless branches.