Poster musings

I presented a poster at Fungal Genetics 2013 way back in March, for which I received a GSA poster award. It was the first poster I’d ever presented at an actual conference, so I thought that was pretty neat. Recently, I’ve been helping one of the undergraduate students in our lab put together a poster about her summer research, which she successfully presented. So I thought I’d briefly share a few tips, tricks, and personal opinions for anyone working on their first poster.

(Before I start, there’s an entire blog devoted to good poster design, which I strongly recommend you browse if you’re interested in making a scientific poster. I read quite a few of the posts before I made my poster.)

With that said, here goes:

Be consistent. When planning your poster, spend some time thinking about your design choices. At a certain point, no one really cares what specific color or font style you choose for the theme your poster. Undoubtedly there are requirements that put restrictions on these things (eg a poster size of 42″ x 36″ will prohibit you from having all of your text at 80pt bold). But this is ultimately the part where you get to be creative. Having said that, consistency is incredibly important. If the header for your “Introduction” section is bright red, italicized, and comic sans font, fine. But make sure the headers for your “Methods”, “Results”, and “Conclusions” sections aren’t green, bold, and Times New Roman, respectively. Granted this is an extreme example, but the more subtle changes are important as well: If you decide at the last minute to change your “References” header from 26pt to 24pt, you had best go back and change the other headers as well. This also goes for your color choices. If you’ve found a pleasing color scheme that works for you (I use Adobe Kuler for this), take some time to write down the RGB and/or CMYK values. That way, if you stop halfway during your poster design process, you won’t have to guess and/or redo your color scheme when you come back to finish it.

“Show grid” and “snap to grid” are your friends. Taking the time to make sure your poster elements are properly aligned elevates the look of your poster and helps to maintain a logical flow of information. Additionally, dividers between sections don’t necessarily need to be physical lines if the text boxes themselves are already aligned. Your audience will be able to mentally pick out the boundaries themselves, and the poster will look more open and less cluttered. But again, be consistent. If there’s a 2″ border on the left side, make sure there’s also 2″ border on the right.

Free can be just as good as expensive. My poster-making tools are LibreOffice, GIMP, and Mendeley. I’m not advocating the use of one program over the other; I’m simply pointing out that these are free alternatives to the not free Microsoft Office suite, Adobe Photoshop, and EndNote. It’s not so important which program you use, just that you know how to use the ones you pick. All of these free programs have a high level of community support, with active development and constant improvements. For example, the cross compatibility between LibreOffice and Microsoft Office grows more robust every day, especially since LibreOffice can open .docx files (MS default) and Office can open .odt files (LibreOffice default). Which brings me to my next tip:

File formats. Know them. There are certain instances when I’ll open a .pptx file (Powerpoint default format) in LibreOffice only to have some fonts be missing and the alignment of text is messed up. If you’re reviewing the poster, this is unfortunate and potentially stressful, especially if it’s very close to the deadline for that poster’s submission. As a poster designer, you can avoid these situations largely by knowing what program your reviewer is likely to be using, and “save as…” accordingly. Alternatively, if you don’t expect or want your reviewer to make formatting changes themselves (eg moving figures around, changing fonts, etc), then you can export or print your poster as a pdf file and submit that for review. This process preserves your formatting changes and has a far greater degree of compatibility than the .ppt or .odp formats.

Finally, autocorrect can work for you. (This works with papers as well as posters.) One of the species I work with is Batrachochytrium dendrobatidis, a chytrid fungus rapidly emerging as a global pathogen of amphibians. I also work with the related non-pathogenic chytrid fungus Homolaphlyctis polyrhiza. I can’t stress enough how little I want to manually type out (and italicize) those two names in full every time I want to reference them. Using initials (eg “B. dendrobatidis“) is obviously acceptable, but doesn’t save all that much time. And while using “Bd” is also acceptable, sometimes it seems a bit informal.
So is there a better way? Indeed there is! Microsoft Word (LibreOffice Writer) has a built in Autocorrect function. The default usage is mainly for sloppy typists and poor spellers, functioning largely to do things like change “teh” into “the”, “aslo” into “also”, and “might of been” into “might have been”. However, you’re able to add your own autocorrect options. So if I want to reference the frog-killing fungus a million times without my wrists seizing up, I set up an autocorrect entry that replaces something like “bd~” with “Batrachochytrium dendrobatidis“. Suddenly, 3 characters become 30 and I still have feeling in my hands. Don’t worry, “B. dendrobatidis” can also be covered by using “b.d~”. I still have to do the italics manually, but this drastically cuts down on the spelling errors and typing time. As long as you’re using shortcut words that won’t appear otherwise, you’re fine.
You can do even more cool things too when it comes to special characters. Here’s a list of special characters that come up quite often, especially in methods sections, and the autocorrect shortcuts I use to automatically insert them:

Character       Shortcut
α                     “a~”
β                     “b~”
γ                     “g~”
δ                     “d~”
ºC                   “^oC”
μl                    “ul”
μm                  “um”
μM                  “um~”

So now I’m covered for my methods section where I talk about the 5 reagents I added or my PCR incubation temperatures, or the entire section about G-protein signaling, without having to hunt for it the “special character” menu every time. If you just use the autocorrect shortcut, you don’t need to break your typing flow to grab it from the special character menu, and you’re covered immediately instead of potentially missing a few spots when you go back through your proofreading.

As I mentioned before, these are just a few things I’ve come across during my short poster-making career. I’m not saying these are the be-all-end-all of poster making tips, and that you will always make the best posters if you do exactly these things, but hopefully they’ll save you some time and frustration the first (or next) time you’re making a poster. If nothing else, the main takeaway messages are to be consistent with your design and to spend a little time learning about the tools you’re using in order to exploit their full potential.

Video games for Science! (pt. 3 and 4)

(Double post since it’s been nearly a year since the last one…)

I’d like to introduce two more “citizen science” / collaborative video games: EteRNA and EyeWire.

EteRNA is focused on folding RNA molecules, in the same vein as FoldIt (a competitive protein folding game described in a previous post). Just like DNA, RNA is composed of four molecules (bases) which are capable of complementary pairing. In DNA, adenine (A) pairs with thymine (T), and cytosine (C) pairs with guanine (G). In RNA, C and G pair exactly the same way, however A pairs with another chemical called uracil (U). So basically a linear sequence of RNA looks similar to a linear sequence of DNA, except RNA has U’s and DNA has T’s.

RNA is typically present as a single strand in the cell, and these strands can fold into three-dimensional shapes based on the above rules. RNA folding, like protein folding, is a complex process. Because the bases on a strand of RNA are free to bind to each other, single-stranded RNA can form so-called “stem-loop” structures, with runs of single strands followed by double-stranded regions with loops. Fully automated RNA folding prediction can be complicated since these (sometimes long-distance) interactions between bases can be hard to capture computationally. Certain 3D RNA structures can have important functions in the cell, so EteRNA works by presenting a specific fold and letting users develop the best sequence under a few constraints.

EyeWire is a new game which attempts to map the connections between neurons in the brain. Understanding these neural connections (apparently known as the “connectome”) will help the creators of EyeWire, the Seung lab at MIT, and the broader scientific community advance neuroscience research. The specific focus of EyeWire is mapping the connections between retinal neurons. These are the cells which help transmit the information from the retinal cells (the ones in our eyes which respond to light) to the brain, where it can be processed and interpreted.

EyeWire accomplishes this goal by presenting users with 2D images of retinal neurons. These 2D images are actually “slices” of a 3D microscope image of a dense collection of neurons. Since a single neuron could potentially be in multiple slices, it’s possible to track it by switching between the slices. Thus, the game is played by visually identifying which neuron slices in a collection of images are in fact part of the same neuron.

Both games have been added to the sidebar, and I’ve separated all of the games into a unique section (since I get the feeling there will be more of these to come).

Video games for Science! (pt. 2)

For those who want or need to write code and haven’t yet settled on a text editor:

Vim Adventures!

This RPG-style video game teaches you how to use the vim text editor, which is incredibly useful (though has a steep learning curve). Vim comes standard on Unix-like systems (so Linux, Mac), so no additional software needed (unless you’re on Windows).

With this game, you can at least try to enjoy learning how to use vim efficiently, instead of being presented with a dry list of commands and definitions like most vim tutorials. At the very least you can play a game that doesn’t take itself too seriously.

This link is also going permanently on the sidebar under Education.

EDIT: Ok. So apparently you can play the first two levels for free, but to access the newer levels as they are released (there are four so far) you have to buy a license. The license is only $5, and the creators say this is cheaper than the full game will be, but I’m still not sure I’m comfortable with charging for an internet tutorial. The money does go toward the concept though, which ultimately is getting people interested/excited in programming (a good thing). In the interest of fairness, I can’t recommend anything other than for you to check out the free parts and support them financially at your own discretion.

Interactive introduction to scale

The Scale of the Universe 2

This could be a very useful classroom tool for an introduction to a wide range of science subjects. I’d say it’s probably elementary school level, and would likely be too trivial to use in a high school setting. But it provides a lot of springboards to other topics, so if you have a few different classroom sessions planned around introducing biology, astronomy, and/or physics, this could be a great framing device.

(I’m adding it to the “education” section on the sidebar. Appropriate copyrights and such are on the page; I did not make that animation.)

Ruby

As I mentioned in the “About” page, the picture at the top of this blog is of my dog, Ruby, looking down on the city of Riverside, CA. It just so happens that Ruby is also the name of a programming language. (It’s actually not coincidental.)

I know that diving (or even cautiously stepping) into a new programming language can be daunting. Online tutorials are available for all major languages, but in my experience, these can sometimes be no more helpful than simply picking up a textbook. There are definitely presentation and design requirements in any tutorial that, if unmet, will make learning the material more difficult. Furthermore, workshops and tutorials require that you provide your own computer with some form of software already installed, which can present its own challenges separate from the fact that you’re learning something foreign.

It’s refreshing, then, to see a tutorial as intuitive and approachable as this one: tryruby.org. The nice thing here is that the tutorial is right there with the interactive coding environment: all you need is a browser. It would be great to see something like this for all languages, so that people with no prior background in computer programming can take these first steps on their own time.

I’m not here to plug any one programming language. In fact, I have a corn snake named Perl. I’m also of the opinion that once the basics are learned in any one language, it’s easy (or at least less difficult) to transition to a more appropriate one. All I’m here to say is that a clean tutorial like this could give someone a basic understanding of programming and at least open the door of the path to a marketable skillset. For that reason, I’ve added this tutorial to the list of Education links.

Video games for Science!

Here is a unique video game that dovetails nicely with protein structure prediction. The game is called Foldit, an online multiplayer game that allows users to help predict protein folding.

Players of this game don’t need to have any prior background in biology or protein science. You could go to that website right now, download the game, and start playing. Teams from all over the world compete to predict the folding patterns of proteins based on a few simple rules: keep the protein compact, keep the hydrophobic (“oily”) residues toward the core of the protein, and make sure that residue side chains aren’t bumping into one another. The interface is fairly straightforward as well: click-drag-drop.

In theory, the concept is similar to any ab initio protein structure prediction program. (That is, a program that predicts protein structure based on only the primary sequence, and not taking into account any available similar structures.) In the fully computational approach, the computer will try to produce the best possible folding pattern. Generally, this is the one that minimizes energy and maximizes the stability of the protein, while following a concrete set of rules (like those above). With Foldit, the players are asked to do the same thing, using a similar set of concrete rules.

It would be impossible for any one lab or researcher to manually predict a protein folding pattern in the “drag-and-drop” way of Foldit; there are too many possible confirmations and the time requirement is too high. Foldit’s strength, therefore, is the fact that the same protein is distributed to many teams across the world.

The stated goals of Foldit are twofold. Most importantly, they want to help predict the structures of medically or economically relevant proteins. These structures will in turn inform drug design and possibly novel enzyme design. The production of biofuels, for example, could be made more efficient if better proteins could be designed. Additionally, the researchers are interested in whether or not human pattern-recognition skills can be useful in this regard: Will coupling computational prediction with manual puzzle-solving make structure prediction more efficient? If so, would it be possible to design a program that implements the strategies that the human players have come up with? These are the questions that this project is addressing.

More recent protein structure prediction approaches blur the line between traditional “homology modeling” (use of existing templates) and “ab initio” (purely primary sequence and energy-based) by using solved structures to inform the energy-minimizing scoring functions. A recent paper about Foldit’s success hilites this methodology.

I’ve added the link to Foldit on the side bar under “Education”. Download it and give it a try. This way, you’re playing games for science, instead of watering crops or flinging birds at pigs or whatever.

Squid in Space (pt. 2)

The space shuttle Atlantis returned to Earth yesterday, completing its final flight and wrapping up the US Space Shuttle program. I was able to get down to Florida to watch the launch on July 8th. Like the previous mission, we sent squid up on this one as well. The samples will be examined like last time and will help add to our knowledge about how bacterial-animal relationships are influenced by weightlessness.

I have some opinions on the completion of the US Space Shuttle program, but this is not an opinion blog. I will say this: the Space Shuttle was one of the coolest, most technologically advanced vehicle the nation has ever built. But times change and all good things must end. And, since this is a science blog, I think the science is the important thing here. I look forward to what we can learn from working and living in space, and the experiments we can perform there. It is in our best interests for NASA to gracefully step out of the taxi/delivery business and realign its focus to cutting edge science and research. Now that the International Space Station has been completed, this research can continue in earnest. I am optimistic about the future of human spaceflight.

For educational purposes, there is now a link on the sidebar that points to NASA’s Education Materials that teachers can use in their classrooms. The grades represented here are K through college, so people should be able to find what they’re looking for. If not, the site is fairly easy to navigate.

Protein Structure Prediction II – Folding

(This is a continuation of a presentation I gave for one of my courses. I’m posting it because it’s simple and I need to digest the Indoor Air 2011 conference I just got back from. If you need to, read the Protein Structure Prediction I post for an introduction to the chemistry of proteins.)

Ok so now the important part: why do proteins fold? And more specifically, why should you care? It is certainly simple and useful to deal with the “linear string” concept: this structure can be easily captured in a text file (which we call “FASTA file”) and we can do sequence alignments with it using BLAST. But the cell is not a text editor, and in the cell, proteins exist in 3 dimensions. The 3-dimensional positions of the various side chains, much more than simply their order, are what really dictate protein function.

Proper folding is therefore essential to proper function and misfolding has significant consequences. Proteins are machines within the cell that control all of the things that keep us all alive. (This is in bold because this is why you should care about protein structure.) The complex biological processes that dictate life as we know it (including environment sensing, cell growth and development, and a whole mess of metabolic reactions) all occur in large part because of proteins. More specifically, they occur because these proteins have a certain shape and folding pattern, and any changes in that pattern can be devastating. (Diseases like Alzheimer’s, Scurvy, Cystic fibrosis, and Creutzfeldt-Jakob, among others, are related to improper protein folding.) So it’s a good idea to understand not only how these proteins fold correctly (because for the vast majority of us, they do) but also what their final structure is, how it interacts with other proteins, and why, specifically, that’s important.

Here I’ll introduce arguably the most important website related to this subject: the PDB. The Worldwide Protein Data Bank is a collection of organizations in the United States, Europe, and Japan. It serves as a repository for all experimentally determined protein structures; that is, proteins that have been examined in the lab using X-ray crystallography or NMR, rather than predicted computationally. This is important because it is a single archive that is both freely and publicly accessible, and manually curated. (It also serves as the main place to get benchmark datasets when developing protein prediction programs.) The three international websites simply function as different portals to the same content; the United States site is www.pdb.org. They have a great PDB-101 section that does an excellent job of explaining the concepts of and providing educational resources for structural biology. I’ve included a link to it under “Education” on the right side of the page. I’d suggest taking a look at it.

In the final section next time, I’ll go over the main types of modelling approaches and provide information on commonly used software.

Updates

Two quick updates:

The three squidonauts (Ulises, Kraken and Penny) shot into space last month have been successfully recovered. They are currently undergoing transmission electron microscopy procedures at the University of Florida in Gainesville, FL. Transmission electron microscopy, or TEM, will allow us to generate very high resolution and detailed images of the light sac post-infection. The results of this analysis should be back next month in time for the final shuttle launch ever. We are planning to launch additional squid then as well.

Also, and unrelated: I’m attending the Indoor Air 2011 conference in Austin, TX. The main focus of the conference is trying to understand and improve the air quality of the indoor environment. This includes any and all structures that humans have built and inhabit: homes, offices, airplanes, etc. The majority of the week-long conference is devoted to chemistry and engineering practices. However, there is a two-day symposium starting Wednesday on the Microbiome of the Built Environment, which is far more relevant to my interests and field of study. The built environment represents a relatively new environment in which microbial communities can be found and studied. (“New” in the sense that we’re just beginning to examine the organismal communities located therein; obviously human-inhabited structures have been around for nearly as long as humans themselves have.) Interested? Check out this website for some additional information.

Update: Indoor Air 2011 tweets can be found at #indoorair2011. I’ll do what I can here.

Protein Structure Prediction I – All about proteins

One of the projects I’m working on is a structural examination of fungal sensing proteins (light-sensing, in this case). So it would certainly be beneficial to discuss computational prediction of protein structure. To that end, I’m reproducing a 20 minute presentation I gave for one of my courses. I’ve divided it into three segments, so continue watching this space.

First, it’s important to understand a little about how proteins work and why protein folding is important. Proteins are composed of amino acids, which are simply chemical compounds with a specific structure (depicted below). Amino acids found in organisms have three main functional groups: the amine group (N), the carboxylic acid group (C), and the side chain (R). In the cell, there are 20 unique, biologically relevant amino acids, each of which differ in the chemistry of their R groups. Each amino acid is encoded by a nucleotide sequence of three letters, which is referred to as a “codon”. Linking the amino acid to its appropriate codon is an RNA molecule known as tRNA. So, when the enzyme responsible for translation, the ribosome, scans the mRNA transcript (that is, the RNA molecule produced from genomic DNA), a specific tRNA is brought in and matched to a specific codon on the mRNA. The amino acid is released from the tRNA and added to the existing chain of amino acids by forming a bond between its N-group and the unbound C-group of the chain. This process continues until the ribosome reaches one of three special codons known as “stop codons”. The end result is a long chain of amino acids, where one end contains an unbound amine group, the “N-terminus”, and the other end contains an unbound carboxylic acid group, the “C-terminus”. If you ignore the R groups for a moment, you’ll see that all amino acids share the same chemical structure. Stripping all R groups from our chain of amino acids produces a “backbone”; simply, the repeating chain of “N-C-C”. So the backbone essentially forms a scaffold onto which side chains are packed, and the chemistry and orientation of the side chains are what give the protein its function.

A generic amino acid

Protein structure comes in four flavors: primary, secondary, tertiary, and quaternary:

Primary structure refers to the linear sequence of amino acids, which can be obtained by using the genetic code. Primary structure is useful when considering sequence-based homology searches (BLAST, for example).

Secondary structure refers to the local substructure within a protein. There are two defined secondary structures: the alpha helix and the beta strand. (Regions that aren’t part of alpha helices or beta strands are referred to as loops.) These structures are defined by the hydrogen bonding patterns between the oxygen on the carboxylic acid group and the hydrogen on the amine group of two different proteins. One complete protein may have several of any of these regions.

Tertiary structure refers to the overall 3D structure of the protein; it captures the number and positions of all helices, strands, and loops in the protein.

Quaternary structure refers to the interaction between subunits in a multimeric protein complex. This level is important since in the cell, many proteins function as complexes composed of smaller subunits. As an example, this is a membrane protein found in E. coli. The gene codes for one single subunit, of which there are 7 in the final protein complex:

Small-conductance mechanosensitive channel

Next time, I’ll discuss why folding is important and how we can computationally predict it.