Genome versus Epigenome

For those of you who care (and those of you who don't, I suppose), I have surpassed the word count goal of 50,000 and am now in the realm of "As long as I finish my plot, I will be happy." Which is a much more laid back and calm place to be than those people who are at 20,000, or 15,000 or (eek!) 5,000 words right now.

And, since word count no longer is of primary importance, I can talk about something else. Namely, this past week's New York Times Science Section, which focused on a subject close to my heart: epigenetics. (Epigenetics and fiction, I just need to mention circus stuff in this post and it will be a summary of my life.)

If you haven't read it, you should -- it's a pretty good read and a decent introduction to epigenetics. Here's a summary article.

And here is my caveat, in which I say that I have a bias. I am a geneticist. I study epigenetics. And, reading this article, I felt the need to stand up for the power of the gene, or at least the DNA sequence.

Robert Tjian gave a talk at Stanford two weeks ago. He's the next President of HHMI, and studies how genes are regulated in a cell-specific manner. And he said, something like three times in his talk and in a smaller discussion afterward, that it all comes down to transcription factors, specific cis-acting binding sites, and, effectively, sequence. Sequence is still everything, sequence is dictating the changes, even when the changes result in variable interpretation of sequence. I asked him about epigenetics, in part because anyone who studies cell differentiation in this day and age has to consider epigenetics, and what he said was, I felt, particularly revealing. He didn't deny that methylation marks, histone modifications, nucleosome positioning, and whatever other epigenetic mark you can think of are certainly important in development and in differentiation, especially in producing stably differentiated cells. However, he said, even these marks, eventually, come down to sequence.

Take, for example, something that seems overtly epigenetic. A cell divides, and upon division, whichever daughter cell is closer to an external signal remains a stem cell while the daughter cell that is farther away differentiates. How does this come down to sequence? Well, there are specific genes for specific receptors on the surface of the cell which interact with that external signal and cause a signalling cascade resulting in a set of transcription factors being upregulated to cause the differentiation, for one thing. And there are other specific genes which organize and create the mitotic spindle such that when the cell divides, one of them actually is closer to the external signal than the other. Even what appears to be an epigenetic effect based on the location of the cell relative to other cells is, in the end, controlled by genes. There are protein-coding genes which you can knock out which disrupt many, many parts of this process, and each causes a defect in differentiation.

Likewise, consider "position effects" in eukaryotic molecular cloning. Depending on where a researcher inserts an ectopic sequence into the genome of a eukaryote (for example, yeast, or mammalian cell culture), that same ectopic sequence will be expressed at different levels, perhaps very highly and perhaps not at all. Surely this is proof that, finally, sequence doesn't matter: same sequence, different expression, right? Only it's not the same sequence, if we look at a slightly larger level. Neighboring genomic sequences differentiate one situation from the other, and dropping in large sequence buffers called "insulator elements" to situate the ectopic sequence in a more neutral genomic environment goes a long way toward ameliorating "position effects."

The article I linked from the New York Times says, basically, that since only 1% of sequence codes for protein, 99% of sequence is irrelevant, and hence, sequence is 99% irrelevant. However, I would argue instead that since only 1% of sequence codes for protein, the remaining 99% is where all the really interesting stuff occurs. It still comes down to sequence; we just need to redefine what we mean by sequence.