David Hanson’s robots are by now somewhat familiar faces, including his Einstein robot currently being used as a research tool at Javier Movellan’s Machine Perception Lab at UCSD, and the punk rock conversationalist Joey Chaos. A less familiar face is that of Bina Rothblatt, the blonde at the end of the table in the above photograph. Bina is a robot commissioned by Sirius Satellite Radio inventor Martine Rothblatt to look like her beloved wife. Take that, uncanny valley!

Photographer Timothy Archibald and I worked closely on this project with the idea of creating portraits, and maybe a kind of family portrait, of the Hanson robots. After flying to Texas to shoot Hanson and robots at his home and workshop in Dallas, Texas, Archibald wrote to me.

“Here is a big house in a Texas suburb that looks normal on the outside. On the inside it is robot making company made up of a floating array of 9-12 employees sculpting things, working on the electrical stuff and writing code for software…taking over the living room, den, kitchen, etc. On the upstairs level is where Hanson, his wife and 3 year old live. They they are in month three of this arrangement. There is no down time. People trickle in at 11:00 AM and stay until 1-3 AM everyday including weekends. They are cranking right now, trying to hit deadlines with The Android Portrait of Bina Rothblatt as well as a potential consumer robot called ZENO. Curiously, Hanson’s son is also named Zeno. There is a story on how that came to be, of course…”

To see more photography from this story, check out DISCOVER magazine’s May 2010 issue on newsstands now.

Katherine Batiste of Hanson Robotics working on a computer with “An Android Portrait Of Bina Rothblatt” sits on the table.

Tomorrow marks the 20th anniversary of the launch of the Hubble Space Telescope. I spent ten years of my life working on that magnificent machine, from using observations of a supernova for my PhD, all the way to helping test, calibrate, and eventually use STIS, a camera put on Hubble in 1997.

Last year, I published Ten Things You Don’t Know About Hubble, and I don’t think I can really add much to it here. I also have a lot of new readers since then, so I’ll simply repost it now as my tip o’ the dew shield to the world’s most famous observatory.

Introduction

On April 24, 1990, the Space Shuttle Discovery roared into space, carrying on board a revolution: The Hubble Space Telescope. It was the largest and most sensitive optical-light telescope ever launched into space, and while it suffered initially from a focusing problem, it would soon return some of the most amazing and beautiful astronomical images anyone had ever seen.

Hubble was designed to be periodically upgraded, and even as I write this, astronauts are in the Space Shuttle Atlantis installing two new cameras, fixing two others, and replacing a whole slew of Hubble’s parts. This is the last planned mission, ever, to service the venerable ’scope, so what better time to talk about it?

Plus, it’s arguably the world’s most famous telescope (it’s probably the only one people know by name), and yet I suspect that there are lots of things about it that might surprise you. So I present to you Ten Things You Don’t Know About the Hubble Space Telescope, part of my Ten Things series. I know, my readers are smart, savvy, exceptionally good-looking, and well-versed in things astronomical. Whenever I do a Ten Things post some goofball always claims they knew all ten. But I am extremely close to being 100% positive that no one who reads this blog will know all ten things here (unless they’ve used Hubble themselves). I have one or two big surprises in this one, including some of my own personal interactions with the great observatory!

Ten Things You Don’t Know About Hubble

Genetic studies suggest that genes have a big influence on a child’s reading ability. Twins, for example, tend to share similar reading skills regardless of whether they share the same teacher. On the other hand, other studies have found that the quality of teaching that a child receives also has a big impact on their fluency with the written word. How can we make sense of these apparently conflicting results? Which is more important for a child’s ability to read: the genes they inherit from their parents, or the quality of the teaching they receive?

According to a new study, the answer, perhaps unsurprisingly, is both. Genes do have a strong effect on a child’s reading ability, but good teaching is vital for helping them to realise that potential. In classes with poor teachers, all the kids suffer regardless of the innate abilities bestowed by their genes. In classes with excellent teachers, the true variation between the children becomes clearer and their genetic differences come to the fore. Only with good teaching do children with the greatest natural abilities reach their true potential.

This study demonstrates yet again how tired the “nature versus nurture” debate is. As I wrote about recently in New Scientist, nature and nurture are not conflicting forces, but partners that work together to influence our behaviour.

This latest choreography of genes and environment was decoded by Jeanette Taylor from Florida State University. She studied over 800 pairs of Florida twins in the first and second grades. Of the pairs, 280 are identical twins who share 100% of their DNA, and 526 are non-identical twins who share just 50% of their DNA. These twin studies are commonly used to understand the genetic influences of behaviour. If a trait is strongly affected by genes, then the variation in that trait should be less pronounced in the identical twins than the non-identical ones.

Florida just happens to collects data on the reading skills of its young children, using a test called the Oral Reading Fluency (ORF) test. The twins’ scores told Taylor how good they were at reading, and the improvement in the scores of their classmates told her how good their teachers were. Crunching the numbers, Taylor found that genes influenced around half of the variation in reading scores (47%), while shared environments (like a common household) accounts for 37% and non-shared environments accounted for 16%.

Genes are clearly important, but teaching mattered too. At the highest echelons of teaching quality, genes explained around 70% of the variance in reading scores. At the lowest troughs, they only accounted for around 30%.

Taylor confirmed the effect of teaching quality in a couple of different ways. She took a sample of 42 pairs of identical twins and found that those whose reading skills were below average did indeed have poorer teachers than those with above-average skills. She also looked at 216 pairs of identical twins, where each twin had a different teacher. Among these children, the difference in quality between their teachers strongly predicted the difference in their reading abilities.

These results are somewhat different to previous genetic studies, which found that around 65% of the variation in children’s reading skills can be explained by genetic factors. These same studies have suggested that outside influences, like family and school, are far less important – the genes are at the wheel, and the environment is in the backseat shouting instructions.

But Taylor says that the twins in these earlier studies often came from similar and wealthy backgrounds. If they all get similar educations, that would mask the effect of teaching. So she deliberately set out to recruit twins from a wide variety of ethnic groups and social backgrounds. A third were Hispanic, a third were white, and around a quarter were black. Half of the children came from families that qualified for free lunches on the grounds of low income.

There are many caveats to the study, which Taylor herself lists. The reading improvements of a classroom may reflect the school, students or resources, as well as the quality of teaching. You might see different results if you used different measures of teaching quality (like class observations), or of reading skill. The effect of teaching quality might also be different in higher education, or in richer schools.

Nonetheless, Taylor’s work does demonstrate that poor teaching constricts genetic variation in reading ability so that it never germinates. Only in the light of quality teaching does that variation bloom. Teachers should be pleased with the result, for, as Taylor says, “Reading will not develop optimally in the absence of effective instruction.” Likewise, putting really good teachers into a classroom won’t magically make all the students into literary Jedis, and (contrary to what some parents expect) it won’t benefit all students equally.

I wrote about something similar in my New Scientist piece – a variant of the MAOA gene can lead to aggressive behaviour, but only in people who were raised in abusive environments. Again, the environment sets the stage in which genetic actors can express themselves.

Reference: Science 10.1126/science.1186149

Image: by Tostie14

More on education:

• Gender gap in maths driven by social factors, not biological differences
• Simple writing exercise helps break vicious cycle that holds back black students
• Teaching scientific knowledge doesn’t improve scientific reasoning
• When learning maths, abstract symbols work better than real-world examples

Since I’ve been talking about the possibility of admixture with “archaics” (I’m starting to think the term is a bit too H. sapiens sapiens-centric, is the Neandertal genome turning out to have more ancestral alleles?) I thought I’d point to a paper out in PLoS ONE which reiterates the basic fact that the overwhelming genetic evidence today suggests a massive demographic expansion from an African population within the last 100,000 years. Study after study has supported this contention since the mid-1980s. The question is whether this is the exclusive component of modern human genetic ancestry, which is a somewhat more extreme scenario. In any case, the paper is Formulating a Historical and Demographic Model of Recent Human Evolution Based on Resequencing Data from Noncoding Regions:

Our results support a model in which modern humans left Africa through a single major dispersal event occurring ~60,000 years ago, corresponding to a drastic reduction of ~5 times the effective population size of the ancestral African population of ~13,800 individuals. Subsequently, the ancestors of modern Europeans and East Asians diverged much later, ~22,500 years ago, from the population of ancestral migrants. This late diversification of Eurasians after the African exodus points to the occurrence of a long maturation phase in which the ancestral Eurasian population was not yet diversified.

They took 213 individuals, a little over half from diverse African groups, and the other half split evenly between Europeans and East Asians, and sequenced 20 distinct noncoding autosomal regions of the genome. ~27 kilobases per person. The noncoding part is important because they are trying to look at neutral regions of the genome, not subject to natural selection (this is obviously an approximation, as there is some evidence that even noncoding regions may have some selective value). The variation is what you’d expect, Africans more varied than non-Africans, and the two Eurasian populations are distinct from each other, but less so than either is from the Africans. Lots of statistics ensue, and an “Approximate Bayesian Computation (ABC) analysis.” I’ll cut to the chase, the highest probability model is illustrated in panel A of figure 4. Expansion out of Africa ~60,000 years ago, major bottleneck, a ~40,000 year interregnum where there was a relatively unified Eurasian population genetically, and then a separation between East and West Eurasians ~20,000 years ago.

I’ll stipulate that I haven’t dug deep into the statistics, nor would I really comprehend all the details if I did spend a weekend on it. But I’m rather skeptical of the 40,000 year period of a common Eurasian population. Reading the text where they discuss this finding it seems clear that it was surprising to the authors, and I’m not sure how convinced they are about it either. It is interesting that the second most probable scenario, B, is a simultaneous expansion out of Africa by two different groups which lead to East and West Eurasians. That makes me a little less confident about the details on Eurasian demographic history overall than I’d already been. They do note that some Y chromosomal data imply that all Eurasian populations may have derived from a Central Asian group, and the settlement of Europe ~35,000 years ago may actually indicate population replacement (though it’s pretty clear that many Central Asian groups have been recently heavily admixed with the incursion of Turks from Mongolia in historical time overlain upon a Iranian substrate, and some sequencing of ancient Cro-Magnon mtDNA shows that their haplgroup is still found in many Eurasian, and even New World, populations). Perhaps there is a more complicated story to be told about the replacement of early modern H. sapiens sapiens by later H. sapiens sapiens. I wouldn’t discount it, but one analysis does not push me to consider this at all likely. Additionally, they obviously couldn’t test the “two wave” model Out-of-Africa whereby there was a southern migration which skirted the Indian ocean along with a north wave which pushed into Central Asia; they didn’t have any “southern” Eurasian samples. Also, I do want to make a note of the fact that they had a lot more parameters in their model than I’m mentioning, including migration between the two Eurasian groups.

But let’s jump to the conclusion and highlight a portion which is relevant to what I’ve been discussing on this weblog over the past few days. As I observe above they constructed scenarios with different parameters to see which fit the data best, and one of those parameters was interbreeding with older hominin groups in Eurasia. Here’s what they say in the discussion:

For those historical and demographic parameters that have been previously studied, our co-estimations are in agreement with previous reports, highlighting the general accuracy of our estimates. For example, our estimation of the replacement rate of archaic hominids by modern humans, although indicating that the introgression of archaic material into the gene pool of modern humans has been minimal, did not rule out the presence of minor archaic admixture of other hominids in modern humans in agreement with previous observations…However, it is important to emphasize that our inferences are based on non-coding neutral regions of the genome and that adaptive introgression from archaic to modern humans may have occurred to a greater extent…Indeed, in contrast to neutral alleles, adaptive variants may attain high frequencies by natural selection after minimal genetic introgression. Future studies comparing coding-sequence variation in modern humans and extinct hominids (e.g. Neanderthals) should help to answer this question.

Their models don’t offer any plausible scenarios where more than 1% of the sequence which they analyzed was derived from populations which were not from the recent Out-of-Africa movement. But, they do specifically say that they lose power to ascertain whether there was admixture at levels below 1%. At some point in the medium term future when we have a fair amount of ancient DNA from Neandertals sequenced, as well as a lot of genomes of modern human beings, if we still don’t find any evidence for alleles which have introgressed from other lineages which had long been separated, the time for hedging may be over. But at this point there’s still some wiggle room. What I’m wondering though is how the University of New Mexico group found lots of evidence of introgressed lineages when other groups have not. Granted, they had 10 times as many individuals and more diverse populations, but presumably far less of the genome. If there was admixture which we could detect, in light of the nearly two decades of this sort of stuff, I assumed it would be cases of adaptive introgression. Here a very low level of admixture could still lead to the increase in frequency of a haplotype which bears the hallmarks of having been in a distinct population from H. sapiens sapiens for long periods of time (like haplogroup D for the microcephalin gene). In other words, I assumed that evidence of introgression would be a story of genetics & natural selection and not genomics & admixture. For instance, particular metabolism genes and the like which new Africa populations might have picked up just like they’d eventually develop their own adaptations from mutation or extant variation if they didn’t admix. I guess 614 microsatellites may not count as genomics, but if adaptive introgression on a few select genes was how we’d detect interbreeding between native Eurasian groups and the Africans this not a way I’d assume you could find any evidence of that.

Citation:Laval G, Patin E, Barreiro LB, Quintana-Murci (201). Formulating a Historical and Demographic Model of Recent Human Evolution Based on Resequencing Data from Noncoding Regions PLoS One : 10.1371/journal.pone.0010284