The brain in technicolor

Perhaps nowhere is the intimate relationship between biological structure and function as evident as in the nervous system. The trillions of connections between billions of neurons in the human brain form an intricate and incomprehensibly complex piece of circuitry.

Jeff Lichtman's group at Harvard has developed a trick called Brainbow. This technique has the potential to address a huge problem in neuroscience: what's the wiring diagram of the brain?

The problem is, we can't figure out how the brain works until we have a better wiring diagram. That's right, we still just trying to get a decent schematic.You know this is a big problem because someone has added the suffix "-omics" to it: connectomics. What is connectomics? It's neuroanatomy, with a fancy name that connotes newness and excitement rather than nap time.

We have some idea of the large scale connections of the brain...which parts of the brain that have high levels of connectivity to what other part. But, to atheist-paraphrase Mies van der Rohe, details are important. Most of the methods traditionally used do not offer resolution at the level of the individual neuron, or if they do, you can only look at one or a few neurons at a time.

Biologists often use fluorescent proteins found in nature to label subsets of cells. The most famous of these is GFP, the Green Fluorescent Protein from the jellfyfish Aequorea victoria. The GFP gene can be expressed as a transgene in any organism, either globally, randomly, or in specific tissues. Any cell with the GFP gene turned on glows green under blue light. As you can imagine, making lots of neurons fluorescent green is not much better than not labeling them in terms of sorting out individual connections. It is useful to express it, say, in the dorsal thalamus, and then see all the the targets of axons from the dorsal thalamus throughout the brain. But which cells in the dorsal thalamus are sending axons to which targets? You can't tell, it's a bag of green spaghetti.

Using a genetic trick and post-processing, however, Lichtman's group allows labelling neurons simultaneously with ten virtual colors, using just three different fluorescent proteins. Here's how it works: a transgenic construct is introduced containing 3 genes for fluorescent proteins of 3 different colors: cyan (CFP), yellow (YFP), and red (RFP). The trick is this, from each construct, only one of this will be expressed. This is due to the presence of an enzyme that will process the construct, the details aren't important here (see below).

 
Fig. 1a from Livet et al. The Cre enzyme can effect the construct in one of three ways: no processing (RFP is expressed), excision 1 (YFP is expressed), or excision 2 (CFP). Which of these occurs is random.

Now it gets more cleverer. Mice are made that have three of these constructs in their genome. Which fluorescent protein will be produced is independently random for each construct. So each neuron expresses 3 random FPs. Here's the important bit: by looking at rations of different fluorescence wavelengths, you can tell which ones, allowing you to treat each of the 10 possible combinatorial patterns independently.

Fig. 4a from Livet et al. Different combinations of fluorescent proteins are detected and assigned a unique color.

What they've done is made it so that these constructs will only be active in cells that normally express the gene Thy-1 (a lot of neurons, but by no means all). Now, when you look at a cluster of neurons in the brain, instead of an undifferentiated mass, you can pick out individuals and trace their projections based on color. Each neuron will only mach 10% of the others.

Using computers and fancy rendering software, the structure and contacts of individual neurons can be extracted from the spaghetti bag. This is a big leap forward in the ability to ask high-resolution questions about neuronal wiring. Most importantly, beautiful pictures!

Panels from Fig. 5 of Livet et al. Individual morphologies and contacts of neurons can be extracted from Brainbow-based imaging of the dentate gyrus.

 

Livet, J., Weissman, T.A., Kang, H., Draft, R.W., Lu, J., Bennis, R.A., Sanes, J.R., Lichtman, J.W. (2007). Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system. Nature, 450(7166), 56-62. DOI: 10.1038/nature06293

Technorati Tags: neuroscience,biology

Read this entire post...

A Good Example

After trashing Nature News yesterday, I'm happy to be able to point out much better coverage of human genetics on the same web site. The article is called "Fear in the Genes," and it discusses a recent paper from the Medical College of Virginia called "The development of fears from early adolesence to young adulthood: a multivariate study."

Here, we are not prattling on about "genes for fearing snakes" and other such tripe. Instead, the claims of the paper are explained in plain language, but without a loss of precision. For example, "A tendency toward fearfulness does have genetic underpinnings..." makes clear a genetic component, without making sloppy claims about genes "controlling" fear or other behaviors. It also nicely illustrates the dovetailing of genes with the environment and other extrinsic factors. Nature and nurture are not neatly separable things, they are participants in the same complex process, the ongoing development of the organism, resulting in phenotypes from the simple to the complex.

Technorati Tags: science journalism,human genetics

Read this entire post...

Is there a gene for bad science reporting?

There is something about human genetics that causes journalists, who may otherwise be reasonable, cool-headed people, to enter a loony-land of speculation and bad reasoning.

From Nature News this April 4, we get a good example: 'Ruthlessness Gene' Discovered. Along with pictures of notorious 20th century dictators, the article describes a recent paper by Knafo et al. of the Psychology Department of the Hebrew University in Israel: "Individual differences in allocation of funds in the dictator game associated with length of the arginine vasopressin 1a receptor RS3 promoter region and correlation between RS3 length and hippocampal mRNA."

I haven't read the paper because I don't have access to that journal through our library. However, looking over the abstract and the description of the work in the Nature News piece, I've gleaned the following:

1. Participants play a game in which they are given the opportunity to share some money, anonymously, with someone they can't see or interact with.

2. This game is called the "Dictator Game," though it could just as easily be called the "Decision Game" or the "Splitting Up the Money Game" or the "Sharing Game" or the "Donation Game" or the "Charity Game"... blah blah blah. No one makes any claim that this game has anything to do with political proclivities.

3. There is a statistically significant correlation between how people behave in this game, and which allele (version) they have of a gene, AVPR1a. AVPR1a encodes a receptor for vasopressin, a neuropeptide that plays a role in social behavior in mammals, notably prairie voles. In this case, it is not the receptor itself that is different between people, but regulatory elements nearby that vary; these differences may lead to spatial or temporal differences in where this receptor is made, or how much of it is produced.

Lets start with the broadest criticism, which applies to a large percentage of science reporting on human genetics: Genes do not specify behavioral traits. Genes do not even specify basic features, like arm length. Genes specify the amino acid sequence of proteins, and as mentioned above, regulatory sequences partially control where and when and at what levels these proteins will be produced. Proteins, in turn, participate in the complex networks that underlie all cellular processes: growth, division, motility, physiological properties. These networks are poorly understood. These poorly understood networks, operating in poorly understood ways contribute to poorly understood developmental processes that are influenced by myriad extrinsic factors, and ultimately result in what we would call a phenotype.

In a nutshell, talking about a single genetic polymorphism in the context of a complex phenotype like a (highly artificial, in this case) social behavior is a huge stretch. What the paper says nothing about:

1. How any actual "dictators" (which the author seems to equate with "selfish/mean people") or non-dictatorial leaders perform in this game.

2. The genotype of any dictatorial or more benign leaders at the AVPR1a locus.

This paper is interesting. It shows a correlation (not a causative relationship or mechanism) between how people behave in a highly contrived and artificial socio-economic game, and their AVPR1a genotype. To leap from this finding to implicating individual AVPR1a genotypes in the complex web of social, economic, personal, political, and just plain random events that produce "ruthless" political figures... it's a gross misrepresentation of what genetics can tell us.

But the leaping to unsupportable conclusions doesn't stop at genetics... the author goes on to make historical claim this tells us something about why people become dictators! 

"If that is true, then apparently ruthless dictators may be motivated not by out-and-out greed but by a simple lack of social skills, which leaves them unable to sense what's expected of them."

Yeah, maybe, or maybe they're just a random collection of jerks. There is nothing in this paper or any other work cited in this article that suggests anything either way.

Very interestingly, a press release on the Hebrew University web offers an interpretation of the same work with opposite emphasis: "Are we genetically programmed to be generous? Hebrew University scientists say yes." While this interpretation is facile and wrong for the same broad reasons, it nicely illustrates the malleability of spin (or is it framing?) that can be loaded on to research like this.

Technorati Tags: science journalism,human genetics

Read this entire post...