PNAS Early Edition highlights 20060802

PNAS publish their own This Week In PNAS Early Edition but below the fold are my picks from the latest batch, which include a million underwater moles, liposuction as a resource for tissue surgery, crummy bioweapons made in the US purely for civilian applications, and the problems of verbing nouns or nouning verbs.

[Published: 02-Aug-06 | Permalink | Category: Writ large | Comments]

Rodríguez and coworkers don't turn fat into muscle, not quite. Instead they describe (doi:10.1073/pnas.0604850103) successfully encouraging stem cells from human adipose tissue to differentiate into smooth muscle. Adipose is easier to come by than bone marrow: these cells are from "processed lipoaspirate", meaning it may one day be possible to turn your love handles into a bladder for yourself or twin (or other immunocompatible person). A nice gift for Xmas.
A marine megamole. An open-access paper from Sogin and coworkers (doi:10.1073/pnas.0605127103) discusses the "very ancient and [possibly] nearly inexhaustible source of genomic innovation" in the sea in the form of unexpected diversity in deep water microbial communities. They quote others who have calculated that the marine biosphere may contain as many as a 3 x 10E+29 microbial cells (a million moles, much more than the grains of sand on the world's beaches). Then they use MPSS to show that the diversity is two orders of magnitude greater than some have suggested. They found a sort of long tail of microbial diversity i.e. a rather large number of relatively non-abundant individuals that, in aggregate, is a significant fraction of the total available. Cagily, and most likely correctly, they suggest that conventional techniques underestimate biodiversity in all such surveys, not just marine ones. J.Craig and his gang won't need to suck so much. I like their thought that this is the reservoir of genetic diversity that allows life to bounce back each time. (This is also another nice piece of publicity for 454 sequencing, who are doing very well recently.)
The big story (here, here, here, etc.) is how bird flu isn't frightening any more. At least, that's how it's being spun. Maines and co report (doi: 10.1073/pnas.0605134103) the most blatantly dual-use experiment since this one: they created "reassortment viruses" of the much-hyped H5N1 and an extant human influenza (H3N2) to see what might crop up in the wild as H5N1 undergoes selection for human-to-human transmission. The result in this case was puny viruses that could not reliably transmit via respiratory droplets (unaltered H5N1 can't either, at least in the ferret model used - H3N2 can). You shouldn't sleep safer though: they've hardly covered the combinatorial possibilities of reassortment and haven't touched the potential mutations at all. That's important because some pandemics e.g. the 1918 one result from mutated strains while others (1957 and 1968, this last by H3N2) from reassorted ones. Still, it's nice they didn't create anything nastier than is already available in nature.
Naming names, Farmer and co (doi: 10.1073/pnas.0602173103) detect a probabilistic relationship between the sound of a word and whether it is a noun or a verb. Nouns that are noun-ier are processed by the brain more rapidly than nouns that are less like the archetypal noun; verbalicious verbs are also understood more quickly. "Marble" is more nouny than verby using their metric, so you can get to its meaning quicker than "insect", which is, to its lasting detriment, irretrievably verbier. On the other team, "amuse" is verbier than "ignore". Ambiguous sentences containing words that could be nouns or verbs (like "needs") were parsed more quickly and comprehended more thoroughly when the difficult word was characteristic of its role in that particular sentence i.e. when the word was a more noun-y one and the noun variant was the key to decoding that sentence's meaning. This seems contrary to the linguistic orthodoxy that, notwithstanding onomatopoeia, words are arbitrary symbols or referrents from which meaning cannot be extracted directly (you can't tell what "chicken" means from that arrangement of letters, however inherently funny that word is) but it isn't - nobody is suggesting you can't get some clues from the symbol (trivial proof: chickens are funny birds, right?). I wonder if poets know this stuff intuitively.
Pseudogenes remain pseudo. Gray and co (doi: 10.1073/pnas.0602216103 show that an interesting result from 2003 isn't so interesting after all, and that one putative function for so-called junk DNA (i.e. expressed pseudogenes modulating the orthologous transcript) must be put back in the theory box.

PNAS Early Edition highlights 20060802

Comments

Post a comment