how we should select our graduate students

interesting piece of analysis by orion weiner using ucsf data link here. it is fantastic that someone was able to do this, and it confirmed some of my suspicions that i've had from sitting on the stanford bio dept grad admission committee. i would posit another way of interpreting the results. it isn't that any of the scores of the admitted students (except number of years of research experience and subject gres) are useless, but that at the top programs, like ucsf, the students scores are above where they start being predictive. another case is nih grant scores - if i recall correctly, there is no correlation with grant score and productivity (sum of impact factor points or some such metric), up to a point, estimated around 30% or so. after the 30th percentile, things get predictably worse though... i suspect something similar is going on with our students, where beyond the 30th percentile things would get predictably worse. so, all in all, we may be doing as good a job as we can on the committee. yeah! perhaps we can do it faster next year...

network nonsense continues. an amazing post by Lior Patcher http://liorpachter.wordpress.com/2014/02/12/why-i-read-the-network-nonsense-papers/

http://liorpachter.wordpress.com/2014/02/12/why-i-read-the-network-nonsense-papers/

I always wondered how long this network stuff would go on after meeting Barabasi at a NATO school on complex systems. It just keeps going and going and certainly seems to have lasted far longer than I had thought it would. Part of it, usefully, seems to have merged with the big data tsunami hitting just about everything. Still, though, it seems to me this is mostly about the statistics of artifacts, scale-free... whatever that means. Most biological interactions are probably meaningless, e.g., see the collected works of Sandy Johnson at UCSF on targets of specific transcription factors. So, even if the nodes and edges were all real, it is unclear how informative the statistics of neutral evolution interactions might be in gaining biological insight.

amanda amodeo defends her phd!

a fantastic effort on identifying histones as the molecular basis of the titration mechanism sensing the dna-to-cytoplasm ratio prior up until the mid-blastula transition in frogs.

why we do science : the man who loved numbers

i reread the book 'the man who loved only numbers' by Paul Hoffman over the holidays. it is a wonderful book and a reminder of why we do this. it is so fun and so engrossing. lessons sometimes forgotten in the competitive world of molecular and cell biology, where the competition often is not about who has the ideas, but rather who gathers the most funding and implements fastest. still, it is possible to have a lot of fun solving problems.

i must say though, that the title is just wrong. after reading the book it would be impossible to describe Paul Erdos as only loving numbers. he clearly loved his mother, first and foremost, his collaborators and all children, the 'epsilons as he called them. he just didn't love money, institutional positions or official praise - the usual motivations of many in the higher echelons of our academy and beyond. after his 30's, i don't think he held any official university position, with the duties responsibilities and steady paycheck that this brings. so the image that emerges, is of a man who loved many people and was loved by many. my quibbles with the title aside, this is a highly recommended book.

post-transcriptional regulation ~ monkey on a typewriter?

excerpt from recent paper by Bruce Futcher and colleagues:

"Monkey on a Typewriter

What is the point of having an RNA binding protein that slightly destabilizes hundreds or thousands of mRNAs via a frequently-occurring four or five nucleotide motif? It is natural to suppose that such an RNA binding protein might be used to co-ordinately regulate its targets, and this might be true. However, in the case of Whi3 and its targets, it is not easy to ascribe any overall function to the target mRNAs, and furthermore, despite some recent progress suggesting Whi3 could be regulated by phosphorylation [20], [59] (Cai and Futcher, unpublished), there is no clear idea of when and how Whi3 activity is regulated, if indeed it is regulated at all. Furthermore, other investigations of yeast RNA binding proteins and their targets have likewise found an enormous number of targets for many of the RNA binding proteins, with no readily definable overall function for these targets. Here, as a speculation, or perhaps just as a null hypothesis, we would like to state the “monkey on a typewriter” hypothesis, which is the opposite of the “co-ordinate regulation” hypothesis. One can imagine that each mRNA in the cell would have an optimum average half-life, and of course this would vary gene by gene, according to the function of the gene. But what is the mechanism by which one mRNA comes to have a different half-life from another? We suggest that if there is an array of different RNA binding proteins, and some of these increase half-life and some decrease half-life, and if each RNA binding protein recognizes a short motif of four or five nucleotides, then each mRNA can evolve to contain a set of short motifs targeting it to some subset of the RNA binding proteins such that, overall, the mRNA achieves its optimal half-life. In this hypothesis, the clients of a particular mRNA binding protein need have no common function, and need not be co-ordinately regulated. They are independently seeking their optimum half-life. One could say that the RNA binding protein “regulates” its client’s abundance, but over evolutionary time. The fact that the binding motifs are short is consistent with this hypothesis–the shortness of the motif is evidence that the RNA binding protein is not highly specific. (A monkey on a typewriter will only rarely produce “Hamlet”, but will often produce “GCAU”–particularly if the keyboard has only four letters!) The ARE hypothesis–that mRNAs are destabilized by very short AU-rich elements [16]–is a branch of this hypothesis."

A Wonderful poetic discussion from the always thoughtful Bruce Futcher. The aggregation of multiple weak effects to give rise to mRNA stability and translatability (could that be a word?!?) could explain why so few translational regulators show up in forward genetic screens. The geometric decrease in frequency down to only 4 bp means these regulators are largely binding everywhere, just not in the same amount. So, the effects have to be weak almost by definition. In contrast, TF binding is some much more specific so it can be strong giving over 10-fold increases by the binding of a single protein...