In these slides, I did my best to communicate to the students what is philosophy behind the Unix systems and why they have been so important in the past. The Unix philosophy is in reality an approach to data analysis and programming: I am happy if I have been able to convince the students that, by studying how the first programmers have approached the problem of data analysis, they will be able to learn good programming practices, and avoid mistakes that have already been surpassed many years ago.

I would like to thank my colleague Brandon Invergo and my supervisor Hafid Laayouni for suggestions on how to improve the slides. Enjoy!

You must know that I am a big fan of using small card papers to organize things. I started using CRC cards from the ExtremeProgramming techniques, and now the way I organize my time is similar to the KanBan technique, although I kind of evolved it independently. In simpler words, I have the habit of cutting A4 papers into 8 smaller A6 papers, the size of a post-it, and use them to take note and to plan my projects. If you visit my office, it is full of collections of “A6″ papers everywhere

One day I may prepare a blog post about how I organize my projects with A6 papers. For now, just consider that trello basically allows me to do on a web page what I usually do on paper. Also, trello allows to share workflows with other people on Internet.. For example, I can show you the schedule of the Linux course that I have made:

my trello board for the "Introduction to Linux" course. Click to see it!

So, I used trello to make 5 distinct sets of cards, one for each of the 5 parts that compose the course. In each of this list, I filled some cards to describe the most important topics that I wanted to talk about in that part of the course. I have used some a red color label to highlight which is the most important message to transmit in each of the parts of the course, the “Take-Home” message.

One of the advantages of using cards is that you are forced to be very concise about what you write. Each card must describe a single topic; there is not enough space for more than one or two sentences. This is good, because if forces you to divide everything into the smallest units. If what you are writing doesn’t fit well in a card, it means that you have to split it into two different messages.

Another nice thing about using cards to plan workflows, is that you can easily move cards from one list to another. For example, at a certain point I noticed that I had to dedicate more time to explain the Unix manual and how to get documentation; so I had to move some of the topics to the afternoon, by moving the cards though the list. This is something quite easy to do if you are using cards; but if I was doing the planning on a normal A4 paper, I would have been forced to rewrite everything from the beginning. Cards are a very flexible tool.

So, if you are looking for a tool to organize your workflows, have a look at trello or at the Kanban methodology. Cards are a cool: and there are really a lot of ways to use them, from taking notes during a seminar, to plan your day schedule, to organize your PhD project.

Together, this chapter is a wonderful recapitulation of the previous four. Enjoy!

This is probably the last or the second last session for this book club. I will be away in the Leipzig course for the following two weeks, and then I will also be busy on May. I will maybe make another slideshow on chapter 6 in April, but I can not commit to it.

A genotype space is a representation of all the possible genotypes that can possibly exist, and in which two neighbor points are different only for one single mutation (Hamming distance is 1).

Until now, in the book club slides, I represented it as a matrix:

genotype space represented as a matrix

However, this representation has many flaws… it should be at least a multi-dimensional matrix, since each node should have exactly n neighbors (where n is the length of the genotype), while in a matrix they can have only 4 (or 8 if you count diagonals).

So, a better representation of the genotype space is a graph, like the following:

genotype space represented as a graph

In this graph, the “genotype” of an organism is a chromosome composed by only 5 bases, and in which each base can take only two values. Each node is connected only to the nodes that differ by a single position; for example, “00000″ is connected to “10000″, “01000″, “00100″, “00010″ and “00001″. Thanks to “jts” from Biostar, now I also know that this is an Hamming graph H(5, 2).

Now that we have a representation of the genotype space, we can take any phenotype of our interest, and mark it in the genotype space. For example, imagine that all the genotypes in green correspond to individuals that suffer a congenital disease:

a genotype network. All the green nodes correspond to genotypes that are affected by a congenital disease (for example)

The genotypes in green correspond to what A. Wagner calls “genotype network”, and other authors call “neutral network”. It is a set of genotypes that have the same phenotype, and that are connected by at least one change.

By exploring the topology and structure of a genotype network, we may be able to make some nice observations. For example, how big is the genotype network of a congenital disease, in human populations? Or, how can a population of individuals explore a genotype network?

There are really a lot of questions that come to my mind when looking at these representation. So, it is a good time for me to search on new literature!!

note: I wrote a small python script to generate a Hamming Graph of binary strings. Here it is: https://gist.github.com/1854319

I am very sorry for the people who have not been accepted, but we have received a lot more applications than the places available, and the selection process has had to be very strict. As Katja Nowick, the organizer of the course, said, this is a sign of how much introductory courses to programming for researchers are needed. Hopefully we will be able to repeat the course or other people will organize similar courses in the future.

In case you have not been selected, I would like to give you a few suggestions on how to start to learn Unix/R/Perl skills.

- Are there any other courses for learn Programming oriented to biologists?

I think that the “Unix and Perl Primer for Biologists” course is a very good resource for researchers wishing to learn the basics of the Bash shell and Perl. Their material is easy to read, and explains everything step by step. The authors also wrote a book (which I didn’t read yet), and released some good material on this website.

Another good course that should not be missed is “Software Carpentry for Biologists“. This course covers a wider range of topics than the other, and, more important, dedicates a good effort on explaining what should be the “good practices” for a bioinformatician. Maybe the contents are a bit more advanced than the “Unix and Perl Primer”, although there are classes on the shell. In any case, once you feel a bit confident on your programming skills, you should definitely read all the materials on Software Carpentry, and make sure you have understood everything before starting a research project. There is also an “Advanced Software Carpentry for Bioinformaticians“, by Titus Brown, focused on Python programming.

- Are there other courses on Programming for Evolutionary Biologists, or on Next Generation Sequencing?

Thanks to reddit/bioinformatics, I have found two other courses similar to ours: a “Workshop on Molecular Evolution” from the University of Texas, and a “Computational molecular evolution” Course from EMBO in Greece. Both these courses seem very valid, although I don’t have any direct experience with them.

A good course on Next Generation Sequence is the Angus course, by Titus Brown. Titus Brown is a skilled bioinformatician and programmer, who developed, among other things, libraries such as Pygr and parts of nosetests. The website of the course is full of good documentation and examples, it should be a good place to start.

- where can I get help?

Internet is a good place where to ask for help on programming related questions. The StackOverflow network is the most active community for anything related to Programming and Unix in general. For next generation sequencing analysis, a good place is SeqAnswers. And, for the general bioinformatics question, biostar is of course a nice resource

The most important message of this chapter is that regulatory circuits can suffer a lot of changes, and yet remain functional. For example, some researchers have change up to 600 transcription factors in E.coli, yet it was still able to survive. Or, as another nice example, galactose metabolism is regulated by two completely different transcription factors in S.cerevisiae and C.albicans, yet these two species are not so distant philogenetically.

Another important message of this chapter is the structure of genotype networks of metabolic circuit. I think it can be well represented by this figure taken from [1]. It represents that, in order to find new phenotypes, a genotype network must be robust to changes (all the possible genotypes must be connected), but also be large, so it is able to explore the genotype space.

[1]Ciliberti, S., Martin, O., & Wagner, A. (2007). Innovation and robustness in complex regulatory gene networks Proceedings of the National Academy of Sciences, 104 (34), 13591-13596 DOI: 10.1073/pnas.0705396104

In the second chapter, Wagner discusses the variability of metabolic networks. How do metabolic networks evolve? How many reactions can I remove or add to a metabolic network, without altering its phenotype? How much the phenotype of a metabolic network is robust to changes?

A possible source of confusion in this chapter is the definitions used. The “metabolic network” is the set of all the reactions that an organism can catalyze; while the “genotype network” is the concept defined in the previous chapter. So, this chapter explains how “genotypes networks of metabolic networks” evolve; be careful to not confuse the two terms. The following figure from [1] can clarify the definitions:

1. Matias Rodrigues, J., & Wagner, A. (2009). Evolutionary Plasticity and Innovations in Complex Metabolic Reaction Networks PLoS Computational Biology, 5 (12) DOI: 10.1371/journal.pcbi.1000613

Click on “Continue Reading” to see a resume of this chapter.

What is an evolutionary innovation?

The title of this book is “The Origins of Evolutionary Innovations”. However, what is a Evolutionary Innovation? As a broad definition, an evolutionary innovation is any new trait that introduces something “novel” compared to before. For example, the discover of photosynthesis has been an evolutionary innovation, or the multicellularity has been a revolutionary thing when it appeared. The first chapter lists some nice examples of evolutionary innovations.

Nevertheless, to simplify, we can think that the concepts explained in the first chapters can be applied to any novel phenotype that has never appeared before in evolution. In fact, any phenotype has been an evolutionary innovation when it appeared for the first time. So, in a certain sense, the first chapters of the book describe how new phenotypes can be found, and how a population can explore all the possible genotypes, to find a novel phenotype.

Definition of Genotype and Phenotype

In this book, the author uses different definitions of “genotype” and “phenotype”, depending to the system analyzed. For example, if I want to study how metabolic networks evolve, I can define the genotype as the set of reactions catalyzed by an organism. Or, if I want to study glycosylation, the genotype can be just the list of glycosylation reactions that an organism can catalyze, and the phenotype can be whether the cell expresses glycosylation or not.

Since in our group we are all evolutionary biologists, this fact caused some sensation. It seems too speculative: if I am free to choose which definition of “Genotype” to use and when, I can keep changing that definition until I have found the answers that I like. If I am not satisfied with the results obtained after defining the genotype as the set of reactions expressed in an organism, I can change the definition until I find something interesting. Fortunately, I think that this is just a confusion related to the fact that we have only read the first chapter.. the other chapters clarify the issue well, as they provide nice examples of which definitions of Genotype and Phenotype to take.

Genotype Space and Genotype Network

Most of the people who attended the session were not system biologists, so I dedicated much time to explain what a “Genotype Network” is. A genotype network (also known as Neutral Network in some literature) is the set of all possible genotypes that have the same phenotype. Some definition can also be found in [1]. The concept of Genotype network is a tool that can be used to study how much genetic variability can exist in a population, without changes to the phenotype. If you want to follow the rest of the book club, make sure you have understood the concepts of genotype space and genotype networks before reading the other chapters.

1. Samal A, Matias Rodrigues JF, Jost J, Martin OC, & Wagner A (2010). Genotype networks in metabolic reaction spaces. BMC systems biology, 4 PMID: 20302636

Well, I don’t promise anything, but since I will do the effort of producing some presentations anyway, I will also publish all the slides here in this blog.

This book describes how new phenotypes are discovered in evolution. In the first chapter, it starts by describing some examples of notable phenotypes that have appeared, such as the Urea cycle and the ability to use glucose as a carbon source. But in general, this book is about how any novel phenotype appears in evolution.

It also explains the concepts of genotype space and genotype network, and how much variability can a population of organisms withstand without having changes in a given phenotype. For example, there are far more possible mRNAs than the number of proteins observed, so it seems that any given protein can be produced by more than one mRNA. This means that an organism can withstand many changes to its DNA, without suffering changes to the structure of the protein. What is the role of this variability in evolution?

There is also a nice paper published on the topic today, in Science: Meyer JR et al, Repeatability and Contingency in the Evolution of a Key Innovation in Phage Lambda, Science 2012.

The book club will take place only in my lab, but if you are interested, you can follow the slides and comment on this blog. (or would it be better to discuss it on Twitter? Let’s use the #evol_innov_book tag on twitter). Enjoy!