NGSchool — Summer School in Bioinformatics. Evolution.

German Demidov
Oct 5, 2019 · 11 min read

DISCLAIMER: we initially planned to divide the text into 2 parts (the beginning of the school and all the rest), but we got so much feedback from the alumni so we had to divide it into 3 parts. Here we put the story of NGSchool2017 and some remarks. The next part, as before, will be published in one week. The previous part is here. The next part is here.

In the first part, we hinted at the importance of collaboration for the organization of educational events, but we feel like this needs to be emphasized a little bit more. Since the beginning, a lot about how we organize the event changed — but the main improvement is the fact that we became a team. Yeah, we all are real people, with our plans, ideas, our feelings, and lives outside of the school (sic!). We work together on this project, we put our souls into the project, we spent a lot of our personal time on it, and — we disagree. A lot. Really hard. Sounds harsh, is it though?

It can be. We all have a vision for the school, and even though we agree in principle, the devil is in the details. We have different tastes and focus on different aspects, so reaching a consensus takes time on occasions. Time, nerve cells, grey hair.

What has to be kept in mind in such moments is that the organization of educational events is a very ill-defined problem with many variables. It is not mathematics. Oh, well, it is, but the solution to this problem is “there is no analytical solution”.

That’s why it is so valuable to have all of us working on the project. Instead of an analytical solution, we propose an ensemble of weak classifiers. Each of us has an idea of how to make the school a bit better. But when we pool our opinions into one — that’s where we get the predictive accuracy. And whenever we argue — we all are sure that we argue because we want to make the school better.

Diversity is the key. Tolerance is the key. Friendship is the key.

So this is the first lesson and the most important one from this part.


Sounds cool, yeah? Hehe, we’ve spent quite a bit of time finding the right words. No, seriously, if you did not feel touched, you don’t have a heart. But how the agreement between members can be achieved in practice? How do we actually manage all the discussions and the organization in general? We have a genie to answer your question.

Eugeniusz Tralle, member of NGSchool team since 2018, PhD student, International Institute of Molecular and Cell Biology, Warsaw, Poland

“Through the years of combined experience of doing this and similar events, we started to figure out what does and what does not work for us. One big thing we always had to deal with is the choice of collaboration tools. As our team is spread throughout multiple countries, we needed to make sure our conference calls were as productive as possible.
I think some of you may already know where this is going: Yes, we tried the Agile Scrum project management approach. Let’s say this brought.. mixed results.
On one hand, it allowed for clear task distribution and every call was extremely productive — each of us would go through the list of tasks we’d assigned ourselves to on the previous call, give an update and listen to everyone else. Absolutely no excess fat on any of these conversations. On the other hand, the calls became incredibly boring and each of us had thoughts about quitting that were at least somewhat related to that. Thankfully, because of the other collab tool we’ve been using, all of us managed to hold out a little longer. Yes, everyone and their mother is right: Slack is a decent project management tool. It’s also nice as a WhatsApp/Messenger substitute, which was surprisingly helpful when we just wanted to bitch about our tasks in between the conference calls. I think what ultimately brought down Agile Scrum for us is that at the end of the day we’re a group of friends who do this as a hobby and we needed our communication channels to be as cluttered and messed up as our private relationships. Again, we argue — hard and a lot — but it’s all a necessary part of the process, the crucible in which each next edition of NGSchool is forged. So to answer your question, we do our daily chatting and bi-weekly calls via Slack and do the rest — meeting notes, grant application drafts etc. — through GSuite. Not super inventive, but it works for us.”

Some other issues were common for all the schools and we will describe them here.

Participants’ selection is always a painful process. You do not know the people who apply and you still need to make a registration form, collect the results, evaluate them and select the most suitable applications. We cannot reveal all the details, but for us, it always was starting with portraying the students — which qualities they should have. Then, according to this portrait (such as “The ideal participant should have a basic understanding of Biology, Statistics, Bioinformatics, use the skills later on and not be over-experienced”) we create a registration form. It usually contains 2 parts: one is multiple choice quizzes and one is the motivational letter, divided into several sub-parts. Then the first part is evaluated using simple statistical analysis (such as taking medians and quantile normalization, with the weighting of the questions according to the percentage of the correct answers among all of the participants), and the second part is evaluated manually by several members of the team. Why do we need these quizzes? Well, many schools live without them, only with a motivation letter. But we have found out that human evaluation is really variable between different people who assess the applications. The quiz part was introduced to reduce the coefficient of variation. Also, the quiz usually takes a couple of hours to finish and it shows the level of a person’s motivation sometimes better than the motivational letter. Admittedly, the correlation between quiz results and the motivational assessment is always positive, but not large. Of course, everything is anonymized and reviewers are blinded where possible.
Once we used the website Rosalind as a test and we were pretty satisfied with the results and — we can recommend doing the same for other schools in bioinformatics. The prerequisite was to solve at least 15 problems from Rosalind and to send us a link to the participant’s profile, then the number of solved problems was extracted with the python script.

Logistics. The obvious, but very important thing that has to be kept in mind is the road to the nearest airport. It is not a problem to hire a bus and transport the participants to a beautiful place far away from the civilization — but then every speaker has to be picked up at the airport and also dropped there, and they arrive and depart at different times, and taxi is not always a good option, especially in case of the late arrival. Or language barrier. Same thing may be said about the accommodation for speakers — students can live in a simpler rooms while for senior professors we want to secure comfortable conditions and provide a sense of appreciation for the time and effort they put into teaching.

IT. Fast internet and cloud computing/VMs for computations. Repeat — FAST INTERNET. If you don’t know how to assess it — imagine 40 people downloading 40GB raw experimental data at once and if you think you can ask everyone to download in advance — don’t be naive. In addition, if you want to run hackathon-style session — you can’t be sure what type of data student might need. One IT person is not enough, two is great, three is a luxury! Expect packages being not installed, VM machines going down and containers falling apart. Last year we were submitting bug-fixes pull requests in real-time, so we can run some specific tools. Remember, Bioinformatics software is usually far from production and has very experimental nature, sometimes you need to think fast how to work around a bug or you will spend 1 hour of precious workshop time and risk to discourage student greatly. Briefly — without strong IT support the school in bioinformatics will be a failure. Dot, indentation, bye-bye. It also limits the possible choice of the venue.

But let’s return to the story of the schools…


#NGSchool2017: Warszawa from the Single Cell perspective

We had a short trip to the beautiful historical part of Warsaw after the school

Our second school was about single-cell methods (literally, recently scientists learned how to analyse hundreds of cells one by one, which looked like magic). This technology has just started to emerge a couple of years before and of course both organisers and participants were extremely excited to learn about such data analysis. The topics of the lectures and workshops covered were:

  • introduction to Linux, Bioinformatics & NGS

and material can be found here.

The new thing this year was parallel sessions (some workshops were done in parallel since, e.g., some people showed interest in renewing the knowledge about the basic statistics while others joined the RNA structure probing lecture) and problem-oriented sessions, when during the last days of the school speakers who were working in different fields announced the topics for discussion and participants were free to join any of the sessions, but instead of a lecture format they could discuss the actual issues they had in the chosen topic and all the participants and the speakers responsible for the session could discuss them. We also had 2 BBQ nights and many scientific contacts and collaborations started there!

The dark bioinformatic ritual of grilling sausages

The problems? Yes, there were a few. E.g., it turns out that we had to repeatedly explain that that fish is not vegetarian. The venue planned to accommodate 2 persons in the room for just one (they were also surprised by the number of beds in their room). One of us had to perform a parallel workshop in an actual gym.

Statistical Gym: 2 reps for t-test, 2 reps for the bench press, repeat

So the lesson we learned — when you find a venue for an educational event, even if it looks fine when you come there for the first time, be ready for any kind of absolutely unexpected complications. Well, in hindsight, we could have expected that fish would be considered vegetarian — since Polish people in general like meat so much — but we relied on the venue’s experience in the organisation of such events… So, conclusion: make sure that all the aspects, even the smallest ones, are discussed with the venue and well-described in the agreement. Thankfully, none of these unfortunate events impaired the school in any sense (except giving the organisers a bit of headache).

“I strongly believe that the headache was caused by smth else…” — Alina Frolova, NGSchool team member since 2017.

The group photo of NGSchool2017

But what the participants and speakers thought about the school? Let’s ask them…

Maja Kuzman, #NGSchool team member from 2018, participant in 2017, University of Zagreb, Faculty of Science, Division of Biology, Department of Molecular biology, bioinformatics group:

“I loved the NGSchool2017 event! It helped me a lot during my work in the year that followed and I stayed in touch with many people I met there. This is a true community and an amazing one indeed. As a matter of fact, some of the people became among my best friends and I loved it so much I decided to help them in any way I can so that more people have the opportunity to experience this!”

This was Maja’s opinion on our school 2 years after she joined the team. Since she took part in the school as a participant, we decided to extract hers feedback she gave to us immediately after NGSchool2017 and the result was curious:

“Things that were especially good: Noam Kaplan was a great keynote. Cool subject, great presentation, interesting method. ggplot — very cool lecture, should be repeated again next year. Davis McCarthy had an amazing protocol / lecture that he shared with us — also something to repeat. Functional genome annotation — Marina had the best presentation for me — everything was clear, well paced and well structured. Andrey’s assembly presentation was very interesting and cool but maybe a bit too long.
I really liked that Aliaksei held a part on R project templates — it was useful and cool, although unrelated to the subject (ChIP seq). This would be a good topic — good practice in R. Also, hanging out with everyone was great =) “

Jacek was a speaker at the first NGSchools (’16 and ’17) and then…he moved to Australia. Jacek, come back, we are missing you at NGSchool!

Jacek Marzec, Bioinformatician at University of Melbourne Centre for Cancer Research, Australia:

“I attended the 2016 and 2017 editions of NGSchool, primarily as a speaker, but it was also fabulous opportunity for me to learn new stuff. From a personal point of view, it was a game-changing event for me. The NGS summer school significantly changed the way I perceive bioinformatics in the context of sharing experience and collaborating. In the pre-NGSchool “era” I used to work with rather “conservative” bioinformaticians who were not keen to share data, codes or solutions. During the NGSchool, I realised that it was not the true picture of bioinformatics. It became obvious for me that the open-minded attitude unlocks both professional and interpersonal potential. That is one of the several reasons why I am so grateful to the NGSchool and its organisers.“

Andrey is a bioinformatician from Russia and he knows the way. But back in times, we did not know that he knows the way. Until 2018 — he was a speaker in 2018 too. His impressions of being a speaker at NGSchool2017:

The beautiful review of Andrey has to be here.

Looking forward to telling you the story of NGSchool2018, NGSchool2019 and our future plans in one week!

Best of luck and greetings!

The main author: German
Read, edited, but not necessarily approved by NGSchool TEAM (namely Alina, Gienio, Kasia, Leszek, Maciej, Maja, Sam).
Many thanks to all the people who supported us, took part in the organization of the school, was a speaker or a participant at one of the summer schools or satellite events, sponsored us.

If you are a participant from NGSchool2017 and you want to leave your feedback too — send it to us.

The first part of the story is here. The next is here.

More From Medium

Also tagged Biology

Top on Medium

Ed Yong
Mar 25 · 22 min read

21K

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade