Friday, January 03, 2014

What is a gene?

A gene is all of the DNA elements required in cis for the properly regulated production of a set of RNAs whose sequences overlap in the genome.   
I formulated that definition c. 1990, when I started teaching genetics to graduate students. I think that the course I actually taught was quite different from the plans leading to that formulation, but I remember sitting for several hours in a coffee shop in Newark airport and coming up that definition. This was after the discovery of splicing, transposable elements, remote enhancers, overlapping genes, nested genes, long noncoding RNAs and many short noncoding RNAs, and I imagined discussing literature on each of these topics and its implications for how a gene might be defined. 1990 was before “tweet-length” could be applied, before the discovery of microRNAs and (most significantly) before complete genome sequences and high-throughput data in the style of ENCODE.


I believe this definition has stood the test of time, and that it will continue to provide a useful understanding of what is meant by a gene. 

The fact that it was written to accommodate work that predates complete genome sequences, ChIPseq and whatever methods are developed in the coming years, should be kept in mind as we face hype about new discoveries changing our view of the gene. I predict that later this year some new work will be described as overturning the idea of junk DNA, or the idea of genes as beads on a string, or the notion that genes are merely their coding information, or perhaps all of these. These discoveries will be said to account for the dark matter of the genome and other deep mysteries that were unsolved until now. Faced with that hype, I will link to this post.

In 2014, as part of my plan to write more but shorter posts, I will also report the history of my own understanding of several of the issues that make defining “a gene” problematic.
--------------------
Mark Gerstein almost immediately pointed out that he had published a very similar definition in 2007:

The gene is a union of genomic sequences encoding a coherent set of potentially overlapping functional products.
See PubMed: Pubmed ID 17567988 or 
Gerstein lab: http://archive.gersteinlab.org/papers/e-print/grgenerev/preprint.pdf or 
Genome Biology http://genome.cshlp.org/content/17/6/669.long

Thursday, January 02, 2014

Michael Pollan on plant behavior, good and bad

A friend asked my view, so I read the recent article by Michael Pollan in the New Yorker, "The Intelligent Plant."

Michael Pollan is a very good writer and he picked an interesting topic. Plant behavior is indeed fascinating and he does a good job of fascinating his readers without obviously going far beyond what can be supported. I also think he does justice to the community of plant biologists by presenting people's views in their own words. However, I fear that he may have incited enthusiasm for bad science. A critical point in the article occurs when he points out that the argument is about language.
Many of the scientists in [Gagliano's] audience were just getting used to the ideas of plant “behavior” and “memory” (terms that even Fred Sack said he was willing to accept); using words like “learning” and “intelligence” in plants struck them, in Sack’s words, as “inappropriate” and “just weird.” When I described the experiment to Lincoln Taiz, he suggested the words “habituation” or “desensitization” would be more appropriate than “learning.” Gagliano said that her mimosa paper had been rejected by ten journals: “None of the reviewers had problems with the data.” Instead, they balked at the language she used to describe the data. But she didn’t want to change it. “Unless we use the same language to describe the same behavior”—exhibited by plants and animals—“we can’t compare it,” she said.
I agree that we should use the same language to describe the same behavior, and applying the words 'behavior' and 'learning' to plants make sense to me. That we use these terms (appropriately, I think) for robots and computers points out that they are neutral with respect to mechanism. However, I don't think that 'intelligence' or 'consciousness' would be appropriate for anything described in this article. The prefix 'neuro' refers to neurons or the nervous system and we know for a fact that plants have nothing like neurons. It's pretty clear that multicellularity evolved independently in plants and animals, and there are important differences, so I find it highly unlikely that plant and animal behavior shares underlying mechanisms. Thus I very much doubt that there is “some unifying mechanism across living systems that can process information and learn.” While fundamental processes common to all life are no doubt shared, more sophisticated signaling is unlikely to be the same. Cell walls make it hard to see how information could be possibly be transmitted through synapses, which are specialized points of contact between neurons. On the other hand, plasmodesmata, channels that allow direct but reguated transport between cells, provide plant cells with the potential for mechanisms unavailable to animal cells. Thus, while communication between the parts of a plant is likely to be as sophisticated, if not more sophisticated, than comparable mechanisms in animals, it is very different, and much less well understood. We would do better to appreciate plants on their own terms. I hope that this article leads more young people into the exciting field of plant signaling. I fear that it may do so for the wrong reasons.


Time-Lapse HD Plants following light

Links:

The Intelligent Plant,” by Michael Pollan in the New Yorker. Dec. 23, 2013. Cleve Backster, an obituary in the New York Times Magazine. The best-selling book, “The Secret Life of Plants,” was inspired by Backster’s research.