This year, apart from the “Introduction to Linux” module, I also taught a short module on “Best Practices for programming in bioinformatics”. It was pure fun, I think I never enjoyed so much giving a talk. I explained a part about Version Control, and another about Scrum, and people were really excited about it. To make you understand how much people liked this talk, consider that three persons invited me a beer after that, which for me constitutes the maximum compliment for a talk.

I have uploaded the two slideshow on slideshare. Unfortunately, the best part of the talk was a live demonstration on how I use these practices during my daily work, but at the moment I can not make these example publicly available. However, you should be able to follow the slideshows anyway.

Feel free to have a look at them and make use of them. If you have any comments, you can add them here or to table. Have an happy paper writing day!

click to access the notes.

]]> http://bioinfoblog.it/2012/11/write-it-clearly/feed/ 0 http://bioinfoblog.it/2012/11/a-small-phylogenetic-game/ http://bioinfoblog.it/2012/11/a-small-phylogenetic-game/#comments Tue, 20 Nov 2012 15:37:31 +0000 Giovanni Marco Dall'Olio http://bioinfoblog.it/?p=1155 Continue reading →]]> I wrote a small web game for the “Week of Science 2012″ (Semana de la Ciencia), a science divulgation initiative organized in Spain. I participated to it as a member in the Institut of Biologia Evolutiva of Barcelona, the institution to which I belong to. The game is in Spanish, but I think anybody can understand it without translation. Click on the image to play with it:

The “Phylogenetic Tree” for the “Semana de la Ciencia” in Barcelona.

If the game is not shown correctly, click on this link: IBE phylogenetic game sc2012

In short, we had 15 minutes to explain to a class of college students (from 12 to 18 years old) how to make phylogenetic trees. This is how we organized the time:

• In the first five minutes, we had a short presentation explaining that we all come from a common ancestor, and that our work of evolutionary biologist is to reconstruct the tree of life. We also explained what a phylogenetic tree is, and how we reconstruct it.
• In the next three minutes, we played the first game. This game was quite easy, and was meant to check if the student understood how phylogenetic trees are constructed. During this first game, one volunteer student had to decide where to put a mammal, a bird and a jellyfish in a phylogenetic tree.
• In the next minutes, we played the second game, which was a bit of a trick. Students had to reconstruct the phylogenetic tree of four protists. Have a look at the “Juego 2: protists” to see it. This game was tricky because there it is no way to come with the correct solution. In fact, after letting the students play for a while, we showed them that the only way to know the real phylogenetic tree was to use the DNA sequences. Then we had a few more slides explaining how mutations in DNA sequences can be used to reconstruct the history of changes in evolution.

To make things a bit more entertaining, we also connected a Wii remote to the computer, so the student who played the game had to use it as a mouse. This was fun to set up, and I think I will use a Wii remote in my next talk .

The activity was a bit condensed in 15 minutes, but I think that more or less all the students understood the basic concept. At least, some made questions, and in general, they seemed to like the game. I hope they will at least remember that DNA can be used to study how species have evolved .

If you want to customize the page, the code is available on bitbucket:

• https://bitbucket.org/dalloliogm/semana_ciencia_phylo

This was the first time I programmed something in Javascript, so the code is a terrible mess. There is a lot of code duplication, and a lot of patchy fixes. But as Agile Programmers say, “Code first and Refactore later”. I think I will work on cleaning this code for next year, so if you have any suggestions on how to make it better, please join the repository on bitbucket.

Fortunately  this year, thanks to a meetup group here in Barcelona, I discovered that there are many ways to improve meetings and make them more interesting. The most interesting is the concept of “Gamestorming”, which is based on transforming group meetings into “games”. If instead of inviting people to attend a meeting you ask them to participate to a short game, people are more likely to participate actively and make good contributions.

Most gamestorming techniques involve blackboards and post-its, and ask people to use them to explain their own opinion.  A simple example of a gamestorming meeting would be a planning meeting where the group leader splits a blackboard into three sections, one for listing different “Project Proposals”, and the other two for “Pros” and “Cons” of each project proposal, and asks the participant to fill the blackboard using post-its. If you want to have a good overview of techniques for brainstorming in general, I can recommend you the book “Gamestorming“, by Gray, Brown, Macanufo, from O’Reilly, which I am reading these days.

In any case, I have been thinking about which planning “games” can be adopted in bioinformatics, or by researchers in general. Here is a list of a what I introduced or planned to introduce to my group:

The “post-publication feedback” game

After a paper is published, researchers usually proceed to the next project, and do not speak too much about what has been done. This is unfortunate, because much can be learned by discussing what has gone well or wrong in the paper, and what can be used to make the next publications better. The “post-publication feedback” game is a way of getting feedback from the authors of a paper, without taking them too much time and without boring them.

Participants: This game should be done after one or two months after a paper is published. All the authors of the paper should participate, although other members of the lab can come as well.

Game: In the game, the group leader or the facilitator splits a blackboard into a few sections. You should choose only three or for section titles from the following list (explanation of section titles is given within parenthesis):

• Message of the paper (Can you resume the most important message of this paper, in a sentence? Sometimes, authors of the same paper may disagree on this, and it is interesting to know if it is the case)
• What would you change? (What would you change if you had to start this work from the beginning?)
• What analysis will you keep? (Which analysis/method/type of visualization did you find so useful that you will use it in other projects?)
• What went well?
• What went wrong?
• What’s next? (What is the next thing to do to continue this work?)

After the blackboard has been splitted into section, the facilitator should give three post-its to each person. After giving 5 minutes to everybody to write, each person is asked in turn to post a post-it to the blackboard, in any section. Each person will be given 2 or 3 minutes to explain what he wrote, and then the turn will pass to another person.

After finishing the game, it would be a good habit to copy all the contents of the blackboard and put them into an archive.

Notes: This game must be very quick, and should not last more than half an hour. There are three reasons to keep this short. First, we want to avoid to create discussions between members of the group (your results are all wrong! I could not finish my part because you didn’t give me the materials! etc..). If someone talks for more than 5 minutes, the group leader or the faciliator should interrupt him and proceed to the next person. Second, if people are aware that there it will be no discussion on what they will say, and that they won’t be forced to explain much of what they wrote, they are more likely to give negative comments, which is what we want to get. The third advantage of making this game quick is that people won’t get bored after the meeting, so it will be easier to propose them to do the same after future publications. In theory, members of a group should get the habit of making one such meeting after each publication, and not be depressed by it.

The “please leave me a feedback post-it” poster game

This is a technique that I applied to a poster presented at the ECCB conference this year.

Basically, when you design the poster, leave a section where people can give feedback on the poster. The best thing is to ask three different questions, because this makes it easier for people to reply.

The feedback section of my poster at ECCB 2012

To do this game, it is very important to be gentle and educated. Explain your poster to the people who are interested, as you would do normally; but once you have finished, ask them to leave a comment on the poster, using a post-it.

When I tried this at the ECCB, it had a huge success.. Many people came and left me comments on the poster. Unfortunately, the contents of the poster are still unpublished, and I don’t feel confident publishing it here; but I will write a summary post on this blog once I will be confident enough.

The “What would Sabeti think of this manuscript” game

This is a game that I applied to improve a manuscript and a poster. Pardis Sabeti is the name of a renown iranian population geneticist, and one result I got from the application of this game is an idea of what she would think of a poster I prepared.

Do you know how Albert Einstein had the idea of relativity? He was trying to imagine what would it be like to be a ray of light. He asked himself: if I was a ray of light, how would I see the world around me? Sometimes, good ideas come when you try to imagine the point of view of other people or things.

Participants: you can do this game by yourself, or invite other group members, if they have time. If you can ask somebody else to fill the names and the situations for you, it will be better.

Game: Take a piece of paper, split it into many little pieces, and write one of the following names on each piece:

• Pardis Sabeti
• Kimura
• Stephen Wright
• your boss
• the names of some of your collaborators
• your grandmother

You can substitute Sabeti and the other names with scientists known in your main field of work

Take another paper, split it in many little pieces again, and write some random situations on each:

• at a conference
• before getting to sleep
• after a bicycle ride
• reading papers early in the morning
• at the end of the day, after a very long meeting
• while eating chocolate

Once you have finished, mix the two sets of pieces, and extract combinations of person/situations. For each combination, try to fill a table containing the following columns:

• What is the Principal Message?
• Doubts
• Which section would he/she skip?
• What would he change if he had to do the poster?
• What would he/she do after reading the poster?

Here is an example of what I filled for a poster of mine:

a table I filled for a poster I presented

This game is a simple exercise that will help you imagine how people will react to your work. It will help you fix minor details, and improve the presentation of your work. The fact of combining a person with a situation may seem silly, but it makes it easier to empathize with other peoples’ thoughts. In fact, it is easier to imagine other peoples’ mind if you add some funny element to it.

Other Games

I think I will open a discussion on Biostar to see if somebody else has experience with gamestorming. I know that people at Uniprot use it for getting feedback on their interface, but let’s see if somebody else has something to add.

EDIT: this is the link to the Biostar Discussion on Gamestorming: http://www.biostars.org/post/show/53788/gamestorming-for-bioinformatics . You will find more techniques there.

The pathway of N-Glycosylation can be ideally splitted into two separate parts, one upstream and one downstream of a process known as Calnexin/Calreticulin Cycle, in which an intermediate product of the pathway is involved. In theory, given their function, we can hypothesize that the two parts of the pathway are exposed to different selective constraint, and evolve at different paces among human populations.

The biology and function of the two parts of the pathway are explained in details in the article, but I will try to summarize them here. The upstream part of this pathway is required for this Calnexin/Calreticulin Cycle, a mechanism of folding quality control, so we can expect that all of his genes are conserved among populations. On the other hand, the downstream part of the pathway is involved in host-pathogen interactions, and can be expected to be more variable when comparing populations that adapted to different environments. In the article we have shown that in fact, signatures of population differentiation are more abundant in the downstream part of the pathway.

Unfortunately I don’t have much time to prepare a good presentation to illustrate the paper, but I have uploaded a short resume to slideshare. Have a look at it if you are interested:

You can also check this previous post, where I explained briefly that the main theme of work done in our lab is to study how selective constraints are distributed along the genes of a pathway.

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