Thursday, October 30, 2014

Deconstructing creationist "scientists"

I’ve been fascinated by creationism since I first moved to Tennessee over twelve years ago –home of the Scopes “monkey” trial. And though I’ve been away from Tennessee for about seven years now, creationism still fascinates me. I find it interesting not because their arguments are persuasive or scientifically credible –they’re absolutely not; but rather my interest in it is as a social or maybe psychological phenomenon. Why, in the light of so much compelling evidence, do otherwise intelligent people hold on to something that contradicts the record of life that surrounds us. I’m a biologist because I find the tapestry of life full of wonder and richness, with an amazing story to tell.

But what fascinates me most of all are trained scientists, who hold legitimate PhDs, who take up the cause of creationism. This is interesting from two angles –first the ‘scientists’ (more on them later), and second the organizations that support and fund their operations. Creationist organizations readily adopt and promote these scientist-turned-creationists, even though they routinely belittle and try to undermine working scientists. Its like the Republican party that dismisses the Hollywood elite as not real Americans, but proudly flaunting Chuck Norris or Clint Eastwood. When the PhDs are on the side of creationism, they are great scholars with meaningful expertise, and when they are against creationism (as are 99% of working scientists) they are elitist and part of a conspiracy.

Enter the latest parade of creationist scientists, who’s authority is meant to persuade the public, at a  ‘Origin Summit’ at Michigan State University in a few days. The first thing you see are four bespectacled PhDs, who are authoritized by the fact that they are PhD ‘scientists’. They are: Gerald Bergman, Donald DeYoung, Charles Jackson, and John Sanford. But, unfortunately for them, not all scientists are created equally.

What makes a scientist? That is not easily answered, but education is one element –and having a PhD from a recognized program and University is a good start. But being trained is not enough, there needs to be some sort of evaluation by the broader scientific community. First and foremost, a scientist needs to communicate their research findings to other scientists by publishing papers in PEER-REVIEWED academic publications. Peer-reviewed means that experts on the topic with examine your paper closely, especially the experimental design and analysis, a provide criticisms. All papers are criticized at this stage, but those with especially egregious problems will not be published. Scientists are also evaluated by other scientists when applying for research funds, being considered for promotion (for example, your record and papers should be sent to 5-8 scientists so they can evaluate the meaningfulness of your contributions), or being considered for scientific awards.

Table 1: How to know that you are doing science.

So then, the ability to publish and survive scrutiny is paramount to being a successful scientist. Of course someone who subscribes to science as conspiracy will say: “wait, then scientists control who gets to be a scientists, and so those with new or controversial ideas will be kept out of the club”. The next thing to understand is what makes a scientist “famous” within the scientific community. The most famous scientists of all time have overturned scientific orthodoxy –that is the scientists that were trailblazers and who came up with better explanations of nature. Many scientists appreciate new ideas and new theories, but work on these has to be scientifically robust in terms of methodology and analysis.

Now back to our Origin Summit scientists, how do they compare to normal expectations for a successful scientist? We will use the average expectations for an academic scientist to get tenure as our benchmark (Table 1). First, Gerald Bergman –biologist. He has a staggering number of degrees, some from legitimate institutions (e.g., Wayne State University), and some from unaccredited places with dubious legal standing (e.g., Columbia Pacific University). He had a real faculty position at Bowling Green University but was denied tenure in 1979. He claims that he was fired because of his anti-evolution religious beliefs (his claim –which to me says his creationism cannot be science). He went to court, and long-story-short he lost because he misrepresented his PhD to get a job in the first place. More importantly to our story here is, what was his record? Fortunately for us, scientific publications, like the fossil record, accurately reflect historical events. Looking through scholarly search engines for the period of time between 1976-1980 (when he would be making a case for tenure) I could only find one publication credited to G.R. Bergman, and it appears to be a published version of his dissertation on reducing recidivism among criminal offenders. Published theses are seldom peer reviewed, and this is certainly not biology. He does not meet our basic expectations for the scientific authority he is promoted as.

Next, is Donald DeYoung –astronomer. He is a professor in the Department of Science and Mathematics at Grace College, and Christian post-secondary institution. It has some accreditation, especially for some programs such as counselling and business. Its not fully accredited, but it seems to be a legitimate Christian school. I searched for legitimate peer-reviewed publications, which was tricky because there also exists another D. B. DeYoung, also on the math/astronomy side of the business. If we ignore his non-peer reviewed books, there may be only one legitimate publication from 1975 in the Journal of Chemical Physics, looking at a particular iron isotope –nothing to do with the age of the Earth or evolution. One paper, so he does not meet our expectations.

Third is Charles Jackson with a PhD in education. There is nothing meaningful on this guy to suggest he is a scientist by any stretch of the imagination. Next.

Finally, we have John Sanford, a geneticist. Now we are getting somewhere! How can a person who studies the basic building blocks of life, deny its role in shaping life? He is a plant breeder and was in an experimental agriculture station associated with Cornell University. I found about a dozen real papers published in scientific journals from his pre-tenure time. None are actually on evolution; they seem to be largely about pollen fertilization and transfer, and production of crops. His publications definitely changed later in his tenure, from basic plant breeding to creationist works. Most interestingly, he has a paper on a computer simulator called Mendel’s Accountant, published in 2007, that simulates genetic mutation and population fitness –the basic stuff of evolution, but which can presumably be used to support his theories about mutations causing ‘devolution’ and not the fuel for real evolution. I read the paper. The genetic theory underpinning is not in line with modern theory, and this is further evidenced by the scant referencing of the rich genetics literature. Most of the models and assumptions seem to be made de novo, to suit the simulation platform, instead of the simulator fitting what is actually understood about genetic mechanisms. I assume this is why the paper is not published in a genetics journal, but rather a computer science one, and one that is not listed in the main scientific indexing services (often how we judge a journal to be legitimate). Regardless, of the scientific specifics, Sanford is a legitimate scientist, and he is the one person I would love to ask deep questions about his understanding of the material he talks about.

The one thing to remember is that a PhD does not make one an expert in everything. I have a PhD in ecology and evolution, but I am not competent in basic physiology for example, and would/should not present myself as an authority to a broader public who may not know the difference between phylogeney and physiology.

So, at the end of the day, here is another creationist conference with a panel of scientific experts. One of the four actually deserves to be called that, and even then, he is likely to be talking about material he has not actually published on or researched. There is a reason why creationist organizations have a tough time getting real scientists on board, and instead are relegated to using mostly failed hacks, because there isn’t a scientific underpinning to creationist claims.

Monday, October 27, 2014

Making multi-authored papers work

Collaborative writing is almost unavoidable for ecologists – first author papers are practically a novelty these days, given the dominance of data-sharing, multidisciplinary projects, and large-scale experiments. And frankly, despite the inevitable frustrations of co-authors, collaborative writing tends to make a manuscript better. Co-authors help prevent things from getting too comfortable: too reliant on favourite references, myopic arguments, or slightly inaccurate definitions.

The easiest collaborative writing, I think, involves small numbers of authors. Writing with large groups of people – and for me that’s probably anything over 5 – has unique difficulties and challenges. Collaborative writing with large groups has two types of challenges: first, the problems innate in attempting to find consensus from many competing opinions; second, the logistical constraints and challenges that arise with having many authors attempting to contribute to a single manuscript.

I’ve recently been lead author/wrangler on a manuscript with 15 authors. It seems to be turning out really well, mostly because all of the authors are interested and invested: all 15 have made significant contributions to the text. I’m by no means an expert on the topic of large collaborations, but I wanted to share some of the things I learned (or wish I had known to start with). All of this assumes that the writing process is indeed collaborative; if it is actually one or two main authors and a bunch of non-writing authors this may be much simpler (if prone to its own set of frustrations).

Process: It’s important to determine how things are going to be done early on and keep everyone updated on how that process is going. If parts of the manuscript will be split up, or, if certain figures and analyses will be done by particular people, that should be determined early on and reasonable timelines agreed on. Whoever is managing or leading should keep in touch with all of the authors with updates and timelines, so the project doesn’t fall under the radar. Some thought should really go into what software you will be using, since once you’ve committed it’s difficult to switch. A lot of the time, frankly, you’re limited by the lowest common denominator– programs need to be broadly available (usually free or else very common) and not require a higher level of technical skill than most authors are comfortable with. This is the downside of using LaTex or GitHub, for example. It’s easier (better?) to use an inferior program than to have half of the authors struggle with the learning curve on the program you chose. For that reason, programs that centrally host files, papers, and analysis, like Dropbox, Google Drive or folders hosted on a private server are popular. As with every part of this process, version control, version control, version control. GitHub is the most common choice for version control of software code. Dropbox allows you to revert back to older versions of files, but with limits (unless you’re paying for the pro version, I think).

The more people that are involved, the more variation to expect from your plans: deadlines will be missed, key people will be on holidays, and not everyone will feel the same level of urgency. Note: if you give 10 academics a deadline, 1 person will be early, 7 will finish in the final hours before the deadline, and the rest will want an extension. Consider having explicit deadlines for important milestones, but assume you’ll need to provide some flexibility.

Editing and revising: In the best case scenario, writing with a large number of people is like having an extensive peer review before the paper ever gets published. If you can satisfy each of these experts, the chances of the manuscript making it through peer review unscathed are much higher.

When sending a draft out for edits and revisions from multiple authors it may be helpful to be clear on what you are hoping for from this revision. What should the other authors focus on? Scientific merit, appropriate references, clarity and structure, and/or grammar and style? It may be that any or all opinions are welcome, but getting edits of prose tense or “which” vs. “that” may not be helpful on an early draft.

I’m not sure if there is really a perfect program for collaborative writing/editing that fits the ‘lowest common denominator’ requirement. Optimally a program would be free or very common, require little in the way of installation, allow real time co-authoring, commenting, version-control, and easy import and export. Problems with compatibility between different operating systems, for example, can seem minor with a single user but turn into a nightmare when a document is being opened across many different systems and versions. For smaller papers, I think many academics simply email a copy of the manuscript (often as MS Word or a PDF) around to the authors, and that’s workable for 3 or 4 sets of comments. But dealing with 15 conflicted copies of a manuscript sounds like hell. Using Google Docs/Google Drive was the compromise choice, and it mostly fulfilled our needs, with some irritations. The benefits includes that Google Docs now has different editing modes: 'editing', 'suggesting', and 'viewing'. Only 'editing' allows direct changes to be made to the text. The 'suggesting' mode is more like ‘track changes’ in MS Word, and allows co-authors to comment, add or delete text, in such a way that the main author can later choose to accept or reject each suggestion. The biggest benefits of G.Docs are that co-authors can edit at the same time, in real-time, and so the comments tended to be very conversational, since each co-author can respond to other co-author suggestions. This really helps identify when there is consensus or different opinions among authors. The downside was particularly that some authors prefer being able to edit offline or in general follow the process they are most comfortable with. It seems like restructuring a manuscript is more difficult in shared manuscript, where others might disagree, than on a personal copy. If a few authors dislike collaborative edit, you will still end up with a few conflicting copies, no matter how hard you try to avoid them. There are probably better ways, although I haven’t figured them out yet, and hope someone will comment. For users of LaTex, there is an online collaborative program—writeLaTex—that might be useful. Also, though I’ve never tried it, penflip looks pretty promising as an alternative to Google Docs.

No matter what program you use, you’ll end up with many comments and edits, often conflicting opinions. I think it’s usually good best to defer to the subject matter expert – if a co-author wrote the seminal paper on the topic, consider what they say. That said, without a strong vision, many-authored papers can be unfocused, and trying to make everyone happy almost certainly will make no one happy. After taking into consideration all the comments and expert opinions, in the end the main author has the power :)

Postscript - Authorship order/inclusion/exclusion is always difficult when so many people are involved. Some advice here; also NutNet has some rather well thought out authorship guidelines.

Wednesday, October 15, 2014

Putting invasions into context

How can we better predict invasions?

Ernesto Azzurro, Victor M. Tuset,Antoni Lombarte, Francesc Maynou, Daniel Simberloff,  Ana Rodríguez-Pérez and Ricard V. Solé. External morphology explains the success of biological invasions. Ecology Letters (2014) 17: 1455–1463.

Fridley, J. D. and Sax, D. F. (2014), The imbalance of nature: revisiting a Darwinian framework for invasion biology. Global Ecology and Biogeography, 23: 1157–1166. doi: 10.1111/geb.12221

Active research programs into invasion biology have been ongoing since the 1990s, but their results make clear that while it is sometimes possible to explain invasions post hoc, it is very difficult to predict them. Darwin’s naturalization hypothesis gets so much press in part because it is the first to state the common acknowledgement that the struggle for existence should be strongest amongst closely related species, implying that ‘invasive species must somehow be different than native species to be so successful’. Defining more generally what this means for invasive species in terms of niche space, trait space, or evolutionary history has had at best mixed results. 

A couple of recent papers come to the similar-but rather different-conclusion that predicting invasion success is really about recognizing context. For example, Azurro et al. point out that despite the usual assumption that species’ traits reflect their niches, trait approaches to invasion that focus on the identifying traits associated with invasiveness have not be successful. Certainly invasive species may be more likely to show certain traits, but these are often very weak from a predictive standpoint, since many non-invasive species also have these traits. Morphological approaches may still be useful, but the authors argue that the key is to consider the morphological (trait) space of the invaders in the context of the morphological space used by the resident communities.
Figure 1. From Azurro et al. 2014. A resident community uses morphospace as delimited by the polygon in (b). Invasive species may fill morphospace within the same area occupied by the community (c) or (d)) or may use novel morphospace (e). Invasiveness should be greatest in situation (e). 
The authors use as an illustration, the largest known invasion by fish - the invasion of the Mediterranean Sea after the construction of the Panama Canal, an event known as the ‘Lessepsian migration’. They hypothesize that when a new species entering a community that fills some defined morphospace will face one of 3 scenarios (Figure 1): 1) they will be within the existing morphospace and occupy less morphospace than their closest neighbour; 2) they will be within the existing morphospace but occupy more morphospace than their closest neighbour; or 3) they will occupy novel morphospace compared to the existing community. The prediction being that invasion success should be highest for this third group, for whom competition should be weakest. Their early results are encouraging, if not perfect – 73% of species located outside of the resident morphospace became abundant or dominant in the invaded range. (Figure 2)
Figure 2. From Azurro et al. 2014. Invasion success of fish to the Mediterranean Sea in relation to morphospace, over multiple historical periods. Invasive (red) species tended to exist in novel morphospace compared to the resident community. 
A slightly different approach to invasion context comes from Jason Fridley and Dov Sax, who revision invasion in terms of evolution - the Evolutionary Imbalance Hypothesis (EIH). In the EIH, the context for invasion success is the characteristics of the invaders' home range. If, as Darwin postulated, invasion success is simply the natural expectation of natural selection, then considering the context for natural selection may be informative. 

In particular, the postulates of the EIH are that 1) Evolution is contingent and imperfect, thus species are subject to the constraints of their histories; 2) The degree to which species are ecologically optimized increases as the number of ‘evolutionary experiments’ increases, and with the intensity of competition (“Richer biotas of more potential competitors and those that have experienced a similar set of environmental conditions for a longer period should be more likely to have produced better environmental solutions (adaptations) to any given environmental challenge”); and 3) Similar sets of ecological conditions exist around the world. When these groups are mixed, some species will have higher fitness and possibly be invasive. 

Figure 3. From Fridley and Sax, 2014.
How to apply this rather conversational set of tenets to actual invasion research? A few factors can be considered when quantifying the likelihood of invasion success: “the amount of genetic variation within populations; the amount of time a population or genetic lineage has experienced a given set of environmental conditions; and the intensity of the competitive environment experienced by the population.” In particular, the authors suggest using phylogenetic diversity (PD) as a measure of the evolutionary imbalance between regions. They show for several regions that the maximum PD in a home region is a significant predictor of the likelihood of species from that region becoming invasive. The obvious issue with max PD being used as a predictor is that it is a somewhat imprecise proxy for “evolutionary imbalance” and one that correlates with many other things (including often species richness). Still, the application of evolutionary biology to a problem often considered to be primarily ecological may make for important advances. 
Figure 4. From Fridley and Sax 2014. Likelihood of becoming invasive vs. max PD in the species' native region.

Monday, October 6, 2014

What is ecology’s billion dollar brain?

(*The topic of the billion dollar proposal came up with Florian Hartig (@florianhartig), with whom I had an interesting conversation on the idea*)

Last year, the European Commission awarded 1 billion dollars to a hugely ambitious project to recreate the human brain using supercomputers. If successful, the Human Brain Project would revolutionize neuroscience. (Although skepticism remains as to whether this project is a more of a pipe dream than reasonable goal). For ecology and evolution, where infrastructure costs are relatively low (compared to say, a Large Hadron Collider), 1 billion dollars means that there is essentially no financial limitation on your proposal, so nearly any project, experiment, analysis, dataset, or workforce, is within the realm of possibility. The European Commission call was for a proposal for research to occur over 10 years, meaning that the constraints on project length (usually driven by grant terms and graduate student theses) are low. So if you could write a proposal, upon which there are essentially no constraints at all, what would it be for? (*if you think that 10 years is too limiting for a proper long-term study, feel free to assume you can set up the infrastructure in 10 years and run it for as long as you want).

The first thing I recognized was that in proposing the 'ultimate' ecological project, you're implicitly stating how you think ecology should be done. For example, do you could focus on the most general questions and start from the bottom. If this is the case, it might be most effective to ask a single fundamental question. It would not be unreasonable to propose to measure metabolic rates under standardized conditions for every extent species, and develop a database of parameter values for them. This would be the most complete ecological database ever, that certainly seems like an achievement. 

But perhaps you choose something that is still of general importance but less simplistic, and run a standardized experiment in multiple systems. This has been effective for the NutNet project. Propose to run replicate experiments with top-of-the-line warming arrays on plant communities in every major ecosystem. Done for 10 years, over a reasonably large scale, with data recorded on physiology and important life history events, this might provide some ability to predict how warming temperatures are affecting ecosystems. 

The alternative is embrace ecological complexity (and the ability to deal with complexity that 1 billion dollars offers). Given the analytic power, equipment, and man hours that 1 billion dollars can buy, you could record every single variable--biotic, abiotic, weather--in a particular system (say, a wetland) for every second of every day. If you don’t simply drown in the data you’ve gathered, maybe you can reconstruct that wetland, predict every property from the details. While that may seem a bit extreme, if you are a complexity-fatalist, you start to recognize that even the general experiments are quickly muddied by complexity. Even that simple, general list of species' metabolic parameters quickly spirals into complexity. Does it make sense to use only one set of standardized conditions? After all, conditions that are reasonable for a rainforest tree are meaningless for an ocean shark or a tundra shrub. Do you use the mean condition for each ecosystem as the standard, knowing that species may only interact with the variance or extremes in those conditions (such as desert annuals that bloom after rains, or bacteria that use cyst stages to avoid harsh environments). What about ontogenetic or plastic differences? Intraspecific differences?

It's probably best then to realize that there is no perfect ecological experiment. The interesting thing about the Human Brain project is that neuroscience is more like ecology than many scientific fields - it deals with complex organic systems with emergent properties and great variability. What ecology needs, ever so simplistically, is more data and better models. Maybe, like neuroscience, we should request a supercomputer that could located and incorporate all ecological data ever collected, across fields (natural history, forestry, agronomy, etc) and recognize the connections between that data, based on geography, species, or scale. This could both give us the most sophisticated possible data map, showing where the data gaps exist, and where areas are data-rich and ready for model development. Further, it could (like the Human Brain) begin to develop models for the interconnections between data. 

Without too many billion dollar calls going on, this is only a thought experiment, but I have yet to find someone who had an easy answer for what they would propose to do (ecologically) with 1 billion dollars. Why is it so difficult?