Editor's pick: Random analysis to unmask errors
Sonia van Gilder Cooke mentions calls for journals to replace unpaid peer reviewers with paid research analysts, to correct the incentives that scientists have to publish potentially bad science (16 April, p 38). It is unlikely journal editors will find money for that soon.
But it might be possible to fund paid research inspectors to analyse 1 in 50 published articles selected at random. This is done on many production lines and in other areas of public interest.
Editor's pick: Random analysis to unmask errors
One of my thesis advisers was the chemist , who told me of what I call the Brewer Effect. Say a scientist finds a topic that has never been studied, so does so, but does something wrong. The incorrect result is published because there is no literature to flag up the error.
A second scientist sets out to replicate the result – since it is in the literature only once, a confirmatory study would get published. But of course they don't “confirm” the first result, so they assume they did something wrong and don't submit it for publication. Then a third scientist repeats the experiment, with the same error as the first, and this gets published. Now the wrong result is in the literature twice and all think it is correct.
This idea makes our brane spin
Stuart Clark describes how a superfluid Bose-Einstein state of dark matter particles might explain the streams of dwarf satellite galaxies in polar orbit around the Milky Way and Andromeda galaxies (2 April, p 30). There could be another explanation that avoids invoking dark matter or even modified Newtonian gravity.
M-theory postulates multiple parallel universes (“branes”) existing alongside ours in some higher dimensional space (the “bulk”). Gravity is the only force able to propagate between branes. So any matter within a given brane will experience mutual gravitational interaction with its counterparts in other branes.
Over cosmological timescales, matter will gravitationally collapse around common centres of mass (barycentres) that are shared across branes. At galactic scales, matter within branes will have a range of distributions of orbits and velocities about a given barycentre. The gravitational influence of this mass will appear similar to extended “dark matter haloes” around galaxies.
If gravity propagates weakly between branes, these bulk-spanning matter concentrations could extend further than expected, which may account for the interactions between the Andromeda and the Milky Way galaxies billions of years ago that gave rise to the dwarf satellites.
First class post
Some people think in images, others in concepts but neither can be deemed smarter
Rohini Subash to a suggested connection between absence of a mind's eye and scientific achievement (23 April, p 34)
Neanderthal Y genes no mystery
I was amused by the subtle genetic mechanisms invoked to explain the lack of genes from Neanderthals' Y chromosomes in the human genome, in spite of commonly seeing other Neanderthal genes (16 April, p 15).
There's another possible explanation. History and anthropology both show examples of one tribe of humans treating another as the “enemy”, killing the males and enslaving the females. The enslaved females become pregnant and the offspring are adopted into the tribe. If this was done to Neanderthals, it would explain why many of their genes entered the human genome, but none from the Y chromosome.
Evolution: learning or merely entropy?
Thank you for a thought-provoking piece on the idea that organisms learn from their environments, and this can prepare them for future change (26 March, p 34). I couldn't help but be reminded of previous work on the microevolution of cancer cells and recent work by on the physics of self-replication in RNA molecules. In both cases it is entropy that drives the processes.
In Jeremy England's work, the formation of RNA molecules that can replicate vastly increases the entropy of the system as a whole. The molecules that replicate the fastest also increase the entropy of the system by the greatest amount.
Other work on cancer cells paints a similar picture: an increase in entropy drives the evolution of the cancer. Those cancer cells that dissipate energy most effectively thrive and reproduce and their gene networks become increasingly disregulated, allowing the cancer to adapt to its surroundings.
Ultimately, such “learning” may actually just be a measure of entropy. Those networks that increase entropy by the greatest extent are those that prosper in the system that is being tested.
Migration, vigour and concrete fears
Debora MacKenzie's piece on migration is important and timely, given Donald Trump's electoral strategy of appealing to xenophobia and nationalism (9 April, p 29). I was very happy that she challenged many long-standing assumptions and prejudices that people have about migration; it is rare to see such a full-throated, data-driven argument for open borders in the media. Kudos also to Clare Wilson for her article on ending the war on drugs in the same issue (p 16), for the same reasons.
Migration, vigour and concrete fears
I couldn't help but ponder an analogy between MacKenzie's article on how migrants can enhance the economy and the commentary on how non-native species might – in the long run – strengthen local ecosystems, despite some displacement of the locals (5 September 2015, p 26). Of course, the economy is a form of ecosystem in itself.
Is it just a question of semantics to differentiate between the “xenophobic” view of invasive species causing extinctions, and the “liberal” view of non-native species increasing diversity and thereby improving the resilience of the ecosystem overall? Is it simply sentimental to want to keep things as they are instead of embracing diversity?
Migration, vigour and concrete fears
MacKenzie makes a good case for the economic and other benefits brought by immigrants. But the one thing that immigrants cannot bring is more land. Residents of the countries that are attractive to immigrants, however willing they may be to help, might not relish the thought of land disappearing under concrete to provide housing and infrastructure.
The calculus of a sweet tooth
Hugh Robertson mentions sugar taxes being more effective than “nudges” (Letters, 2 April). The problem with the sugar tax, as with any tax, is that people tend not to think about why a desired object has a certain price, but, simply, whether they want the object enough to pay the price.
Breakfast to keep body clocks on time
Keeping your body clocks synchronised appears to have many health benefits (16 April, p 30). An advantage of eating breakfast may be that it keeps the body clocks synchronised, compared with those who get hungry and start eating at random times in the morning, even though the breakfast eaters consume more calories during the day (26 March, p 39).
Flyaway probe, phone home!
Breakthrough Starshot aims to send thousands of “wafersats”, each with a mass of just a few grams, to Alpha Centauri, more than 41 trillion kilometres away (16 April, p 9). How will they beam pictures and data back to us?
Flyaway probe, phone home!
A British Interplanetary Society team has been pondering this problem for a decade. It produced the Daedalus probe design that calls for a 1 megawatt transmitter with a parabolic antenna several metres across. Even with this power, the group calculated that the receiver would have to be placed far enough out in space for the sun's gravity to focus the signal to give an acceptable signal-to-noise ratio at the receiver.
How many senses do we have?
Reporting experiments giving rats an extra infrared sense, Andy Coghlan lists the five traditional natural senses (19 March, p 22). But we already have very many more than five.
We sense heat, cold, angular velocity, linear acceleration, internal pain (such as headache), relative position of limbs (proprioception), digestive system state (nausea), hunger and thirst and more, all without having to touch anything external or using the other four senses.
Apples + oranges = confusing picture
The “floating point operation” or “flop” is an incredibly bad unit to use to compare the power efficiency of the human brain with computing technology. To do so is like comparing the swimming performance of a shark and that of a dandelion (2 April, p 36).
How long does it take the human brain to perform one 64-bit floating point operation? With my own level of mental arithmetic it would run from “hours” to “forever”. It would take a little less time if fitted with a paper and pencil co-processor.
Haslingden, Lanhcashire, UK
The editor writes:
• While the comparison between brains and computers is somewhat apples-to-oranges, many researchers have estimated the brain's equivalent output to be in the petaflops (100 trillion flops per second). For every 64-bit floating point operation that we consciously work through, our brain is doing a great many more.
Anti-gravitational wave hunt
Physicists have discovered antimatter corresponding to particles of matter. Now that they have found gravitational waves (20 February, p 8), will they start looking for levitational waves?