Monthly Archives: May 2009

What I learned in grad school, part 1

One thing no grad student starting out is ever sure of is how long his doctoral degree will take. That was sure the case with me. Professors in my department often told prospective and first-year graduate students that an “average” degree took five or five and a half years or so. I was always a bit skeptical of these pronouncements, because I never got to see any data. So I never knew how long my Ph.D. would take.

But now that I finished my Ph.D., I can finally answer the question! Completing my Ph.D. program took me 2081 days, or about 5.7 years. How does that compare to the “average” grad student? Unfortunately, that’s a question I can’t answer, because I don’t have access to the proper data, and even if I did, I suspect the answer would vary from institution to institution, from department from department, and from year to year.

But I have been keeping track of my classmates who matriculated to MIT with me back in 2003, and who subsequently qualified as Ph.D. candidates*. Twenty-four of the 39 members of my grad school cohort defended their thesis before I did. Two have already defended after me, and I know six more whose defense is imminent. Two of my cohort left the Ph.D. program shortly after qualifying as Ph.D. candidates, for what I understand to be personal reasons. Of the twenty-four who finished ahead of me, I should note that two of these were in a special degree program, designed to award a doctorate in engineering “practice” after about three years of study instead of the usual five.

And speaking of the “usual” five-year Ph.D., I can also say that in my professors were right! The median time to graduation** for my cohort was 1830 days, or almost exactly 5.0 years. (We won’t know the “average” or mean time to graduation until everyone has defended.) I guess I should give a retrospective kudos to my professors for being well-informed and forthright on this subject.

How long does a Ph.D. take?

This chart shows the progress of my grad school cohort through our Ph.D. programs. It includes all graduate students who started study with me in 2003 and subsequently qualified as a Ph.D. candidate. The red dots indicate participants in a special, shorter degree program. As of today, everyone in the cohort who has not yet defended has been matriculated for 2092 days. (Two students left the program before completing their degree, for what I understand to be personal reasons.)

How does my particular cohort stack up? I don’t have data for other years in my department, or from other departments at MIT. Digging up some old national-level data wasn’t too hard, though. From the looks of it, my cohort is doing very well (or, depending on your perspective, is getting off easy). Nationwide in the early 1990s, only 31% of chemical engineering Ph.D. students graduated after 5 years. Compare that to 50% in my cohort this year. I hope we can keep up the good work.
Continue reading


I’m back – kind of

It’s been quite a few weeks since my last post – but a lot has happened. Most notably, I successfully defended my doctoral thesis and thereby earned a Ph.D. in Chemical Engineering from MIT. I’m also preparing for a cross-country move, and an upcoming wedidng. So posts from me are likely to continue to be irregular. Gone but not forgotten, I hope.

It still hasn’t really sunk in.

Scuffles over biofuels policy

Roger Pielke, Jr., highlights a rift developing between farm state Democratic lawmakers and the Obama Administration’s EPA on biofuels policy.

The EPA seems to be following in the footsteps of the recent California Air Resources Board’s decision to model indirect land-use changes when estimating life-cycle carbon emissions stemming from biofuels production. The argument, in the CARB’s words, goes like this:

Land use change effects occur when the acreage of agricultural production is expanded to support increased biofuel production. Lands in both agricultural and non-agricultural uses may be converted to the cultivation of biofuel crops. Some land use change impacts are indirect or secondary. When biofuel crops are grown on acreage formerly devoted to food and livestock feed production, supplies of the affected food and feed commodities are reduced. These reduced supplies lead to increased prices, which, in turn, stimulate the conversion of non-agricultural lands to agricultural uses. The land conversions may occur both domestically and internationally as trading partners attempt to make up for reduced imports from the United States. The land use change will result in increased GHG emissions from the release of carbon sequestered in soils and land cover vegetation.

Though this rule might sound good, an impressive set of experts, (including my doctoral thesis advisor at MIT), saw right through it in a letter (copied here) they wrote to Governor Schwarzenegger as CARB’s decision loomed. The experts’ strongest argument, in my view is that “indirect” land use changes are not included in the life-cycle GHG emissions analyses of other products or processes, so why single out biofuels? For example, biofuels track crude oil in price because (to an extent) they are substitutable goods. Whether from Somali piracy or from OPEC production cutbacks, oil price increases thus also stimulate, through price-based mechanisms, the conversion of more non-agricultural land to biofuels cropland. Surely then the resultant increase in carbon emissions from this land-use change should be attributed to oil producers, right?

Well, that’s not how EPA or CARB seems to be doing it. I am a bit mystified here: most observers agree that GHG emissions abatements for corn ethanol are marginal, at best (and some critics even suggest they could be negative). Stacking the deck against corn ethanol hardly seems necessary when it’s not a strong GHG abater to begin with. In any case, it will be interesting to see how Democrats, will deal with this controversy developing in their ranks.

UPDATE: I added a title to this post.

Plasmid mutagenesis, the easy way

Douglas Young, working in Alexander Deiters’s group at NC State, describes a new method for introducing targeted mutations into plasmids. Unlike previous site-directed mutagenesis methods, this one requires no restriction enzymes or ligases, and uses only two oligonucleotide primers. The key trick is to put photocleavable tags on some of the primer nucleotides. These tags block polymerase readthrough and prevent unwanted primer hybridizations. After PCR, all UV irradiation is all you need to cleave the tags, letting those single-stranded overhangs hybridize. The resulting nicked plasmid can be transformed directly into E. coli. The authors show that in one step, their method can delete huge sections of plasmid, or, even better, insert completely new sequences up to 17 bp in length, at any arbitrary position in the plasmid.

Very cool!

I have two questions. One is whether the authors have tried doing more than one round mutagenesis/insertion completely in-vitro. After annealing to create the nicked plasmid, a quick enzyme treatment can repair the nicks. Does this give rise to a substrate ready for the next round of mutagenic PCR, with no need for transformation and outgrowth of E. coli?

The other is, how long will it take till this technology is commercialized? I hope that the authors hear from custom oligonucleotide synthesis providers (maybe NEB, Invitrogen, IDT, or others) soon.

Subterranean and suboceanic life

In a series of posts of the next few days months (or weeks years as time permits), I’d like do delve a bit into a topic that has fascinated me for some time: life under and inside the Earth’s crust.

Can life survive in the Earth’s crust, in the tiny pore spaces under miles of granitic rocks? What about under the bottom of the ocean – can eke out a life down there?

The answer to both questions is yes. Microbes – both bacteria and archaea – have been found down there, and trying to understand how they live, and how important and extensive these extreme ecosystems are compared to other parts of the biosphere, is a major goal of modern microbial ecology.

The place to begin our discussion is with an oft-cited 1998 article by William B. Whitman, David C. Coleman, and William J. Wiebe. This masterpiece of the underappreciated art of estimation lays out a rough calculation of the total number and mass of prokaryotic cells on the Earth.

The estimated the population of terrestrial subsurface prokaryotes in three different ways: by extrapolating from the (then) extremely limited measurements of free-floating microbial cell counts observed in deep groundwaters; by assuming reasonable values for average sediment/rock porosities and for the average volume fraction of pore space occupied by microbial cell mass; and by extrapolating from measured microbial cell counts found in “unconsolidated sediments”.

Unconsolidated sediments cover the vast majority of the ocean bottom and about 20% of the terrestrial surface, and so even though the data for microbial cell counts in these ecosystems is fairly good, the 80% of the terrestrial surface covered by other materials (e.g., consolidated sediments or granitic rock) leads to considerable uncertainty.

Back in 1998, these guys estimated that in the oceanic subsurface, there were about 355×1028 microbial cells, comprising about 303 Pg of carbon. They could only give a range of possible values for microbes in the terrestrial subsurface, from 25-250×1028 cells, comprising 22-215 Pg carbon. To put that in perspective, all the plants in world – all the forests, all the trees, all the grass, and all of our crops – contain something like 560 Pg of carbon. So subsurface microbial biomass nearly as much as all the plant-derived biomass on Earth.

And about these microorganisms, their environment, and their way of life, we know next to nothing. But we’re learning more, and over the next few posts I’ll highlight some of the most exciting of the recent discoveries in the field.

UPDATE: I’m still working posts on these topics. Stay tuned but don’t hold your breath.

Carbon sequestration

You can now convert carbon derived from a deceased or living loved one of your choice into a commemorative diamond gemstone. Thanks to marginalrevolution for the pointer.

These diamonds probably have the lowest 14C isotopic depletion (highest levels of 14C) of any macroscopically-sized diamonds in the universe.

The minimal-er genome, part 2

Yesterday I blogged about the J. Craig Venter Institute’s work on identifying and synthesizing a minimally-sized 580,000 nucleotide genome.

Of course, “minimal genome” means different things to different people. A number of players are racing to sequence full-sized, natural, giganucleotide-sized genomes at minimal cost. Today, that race gets a little more crowded, with sequencing start-up NABsys announcing that they’ve received a $4 million infusion of venture capital. NABsys’s technique, unlike competitors 454 Life Sciences or Helicos Biosciences, doesn’t rely on fluorescent or luminescent light detection, and thus doesn’t need any type of DNA polymerization or strand extension reactions to generate sequence data.

No light? No sequencing by synthesis? So how do they do it? It’s simple! All you need is nanopores and “moving window sequencing by hybridization”. Well, perhaps “simple” was an overstatement. But the explanatory video put out by NABsys explains things beautifully.

I would like to have seen some references to published literature on the NABsys web site, but even more, I want to know when I can have my genome sequenced for the low low (minimal?) price of $999!