An index to this series appears at the first post.
October 15, 2018
I awoke before my alarm sounded at 6:30 AM. The first-day nerves must have been jangling! At 8:00 AM, Dr. Alexander Kwarteng picked me up from the guest house lobby. Happily, I was all set before time, and off we ran to the Department of Biochemistry and Biotechnology at Kwame Nkrumah University of Science and Technology. An impromptu meeting with Head of Department Prof. Peter Twumasi revealed his excitement for the topic of bioinformatics. He was blown away by the strong response to our week-long program. “Everyone in biology needs skills in bioinformatics,” he said. I couldn’t have said it better!
Head of Department, Prof. Peter Twumasi
As showtime approached (I will be lecturing each morning at 9:00 AM with a lecture split to two segments), I paused for some tea in the departmental lounge, and I took a monster-sized mug to the lecture hall. I learned that I’d be presenting from a raised dais to a room full of lab benches (complete with gas spigots). The students gave me guarded smiles as I came in the room. I was grateful that my friend Samuel Ahuno was willing to take some photographs of the event.
Samuel Ahuno has been a regular participant in DARA programming workshops throughout Africa.
It is always hard to aim at the right level of scientific background when preparing a tutorial for an audience of unknown composition. I built the five talks I’ll deliver from six of eight lectures that I created for B.Sc. Honours students in my home division at Stellenbosch University (the students have completed B.Sc. degrees spanning three rather than four years). As the day progressed, though, I learned that the students ranged from second-year undergraduates to graduate students. In the younger years, these students are accustomed to coursework rather than research benchwork.
I decided to spend more time on each slide (I had included 28 of them). When I spotted a vocabulary term in the title, I tried to take the time to ask the students question to pin down those terms. I know that the material was pretty rough going for some of the students, though, and I hope that they will be okay knowing that more of it will make sense in time. At lunch, Dr. Kwarteng suggested that future versions of this course could start at the most basic point: “What is bioinformatics?” While I’ve made slides to that question before, I don’t feel that they really advance student understanding as much as I would like. I am encouraged, though, that the department feels that this course is something to repeat in later years!
We had encountered a problem even before lunch that occupied my mind as we ate. The wifi reception in the lecture hall was pretty weak, and even when I could get a steady eduroam connection in place, webservers in Europe and North America responded with a lot of latency. Dr. Kwarteng had some tricks up his sleeve, but my practical exercises depended on the use of several web-servers. I began working toward a backup plan, one in which I would teach the students the basics of the R programming language. As we came back to the classroom, we began attempting to connect all the students to the web using KNUST wifi services as well as a cellular modem and router that Dr. Kwarteng had charged up for the occasion. Ultimately, many of the students were online, but when they tried to use the network all at once, the network services blinked out. I activated the fallback. With whatever machines could still use the web, we downloaded the R or R Studio installer. We copied the installer to USB flash drives. We then swapped those drives around the room until everybody had R in place.
I drew a deep breath, and asked, “what language is the best choice for getting started in bioinformatics?” In the remaining 45 minutes, we began covering some basics, such as the assignment and addition operator, some differences between numeric and text variable types, the use of vectors, and the use of factors. Tomorrow I hope to have a tighter organization in mind!
Dr. Kwarteng took me by a shopping strip on the way back to the guest house to acquire a box of cereal, milk, bread, honey, local bananas, and peanut butter. I have a wealth of options!
October 16, 2018
To lead a day of tutorials and then throw all remaining energy into being a tourist can be a bit of a challenge! Today I introduced the students to the informatics required to assemble and annotate a novel genome. It’s a complex lecture, with a lot of moving parts, but the students seemed to like this one better than yesterday’s topic: genome-wide association studies and nsSNV (gene variant) characterization.
If you give me a mobile microphone, I will wander with it!
I felt better prepared to deliver some basics in R, as well. I followed the format of a blog post that covers the essentials well. We still had some time available at the end, so I reused some code I had on hand to discuss the differences between males and females and between South Africans and American in height (all the demonstration code is available from my Google Drive). I felt that the demo went pretty well, but I think that the modeling aspect of the height comparison was a step too far for many of the students. (Rather than ask whether two cohorts have different average values, we were asking whether our four cohorts showed a difference either on sex or on nationality.) I need to ensure that the students receive the information in advance for tomorrow’s demo in gene expression; it will be easier for them to follow along, if they have all the materials in hand.
October 17, 2018
On a day when I had some tourism excitement planned, I needed to focus on my teaching. When I framed my lecture on gene expression for a bioinformatics tutorial in the University of Malawi, I greatly strengthened some sections of the talk that seemed underwhelming before. The challenge this introduced, however, was that my number of slides was rising with each of the first three days of the Week of Bioinformatics (28 on Monday, 35 on Tuesday, and 49(!) on Wednesday). The students seemed grateful for the intermission when it arrived, and yet they seemed focused all the way through the section on finding differential genes and multiple testing correction. I call the morning lecture a success! I celebrated by eating some lovely jollof with chicken takeaway from a restaurant just outside KNUST.
The afternoon tutorial in the R language gave mixed results. I started out with very approachable material. If we run the Welch T Test on data with known mean difference and known variance, how frequently do we reject the null hypothesis on ten trials with three replicates, with five replicates, with seven replicates, and with nine replicates? I think the students felt a little unhappy at the notion that their ability to get a p-value below 0.05 could be a matter of chance. The second part of the R tutorial had the students loading two text files from the disk with the “read.table()” function, and predictably the chief challenge was setting the working directly successfully! So far, so good.
I included the third part of the tutorial because I wanted the students to see what a fuller program looks like, not just four or five lines. I also liked the fact that many of the elements I had discussed in the lecture were in my script, showing them how to implement these ideas programmatically. Helpfully, my simulation of a gene expression data analysis required no external libraries, so we were able to execute the code all the way through to producing a volcano plot (using base graphics). Understanding the code, however, was hardly a success. Rather than following my usual instincts to compute the T tests and log fold changes inside a for loop that runs through all the rows, one by one, I opted to use the “apply” function in R. The code is arguably more in line with what a real R programmer would do, but it is considerably harder to understand for a novice. My students were dumbfounded by seeing the full script, and yet I am glad that they have this code to examine if they need to write something of greater complexity in the future. I must remember to explain to them that all programmers get stuck from time to time, and we all have our little ways to find help!
October 18, 2018
[On the night of the 17th, I injured my knees and ankles. I chronicled the accident in a separate blog post.]
For the first time in a long time, I slept solely lying on my back. Putting one leg atop the other seemed like a recipe for pain, and getting biofluids from my nasty knees on the sheets worried me, too. When the birds woke me in the morning, I still had a few moments before my alarm went off. I am lucky that I got an early start because Dr. Kwarteng came to check on me about a half hour before I expected pickup. I was able to limp stiffly to his car, and Kobla “Atsu” Amewu graciously carried my backpack for me. When I settled onto the couch in Alex’s office, Samuel Ahuno ran to the faculty lounge to get me my extra-tall cup of tea.
Atsu Amewu worked tirelessly behind the scenes to make the Week of Bioinformatics a success.
Together, we talked about how to adjust the course now that I was no longer at 100%. We decided on a relatively minor change. I had not planned out content for the Friday afternoon R tutorial, so we decided to cancel that. As a result, some students attending the workshop from Cape Coast and from Accra might be able to get home a day earlier.
Because relatively few students were in the room at the official 9AM start time, I decided to talk through some of the issues associated with becoming an international graduate student outside their home nation. The students had become very interested in an international graduate student opportunity offered by DARA Big Data (Development in Africa with Radio Astronomy). We hope to have a new round of M.Sc. bursaries available in early 2019. Of course, living away from one’s home country can be daunting for many young people, and being able to afford this option without outside funding is quite a challenge.
I spoke of the value of establishing communications with a particular professor, explaining that it requires that they understand what the researcher is trying to accomplish and discerning how their skill set will be useful to those goals (in order to qualify for a research grant funding their efforts). I also talked about the delights of SAQA, a ratings body for South Africa that evaluates whether a degree earned in a foreign nation is equivalent to the degree of the same name granted by South African universities. I hope that these students are able to benefit from this information!
I had scheduled my favorite topic, proteome informatics (identification, quantitation, and structure), for this Thursday morning. Although I had slightly fewer slides than for Wednesday, I ran even longer. The students were lucky that I was able to wrap up by 11:45 AM. Perhaps they were humoring me because I was obviously hindered by my body. This is the first day that they saw me wearing anything other than slacks and dress shoes. To move all the way to shorts and flip-flops was a bit of a surprise!
Jollof with pulled chicken from a restaurant just outside the KNUST boundary
For lunch, I ate the leftovers from yesterday’s lunch, a yummy jollof. As Atsu had commented, yesterday’s jollof is frequently even tastier than our first experience!
As I munched, I decided that I should revise my afternoon R tutorial a bit. The gene expression example from yesterday had left the students flummoxed, I think largely because I used the R-standard “apply” approach rather than writing it out as a loop. I decided we would revisit that code before we talked about principal components analysis. During the afternoon, I saw that “unrolling” the apply calls into a “for” loop made a lot of the elements clearer, even if the code was quite verbose afterwards. I talked about creating a vector of “NAs” in which to store the p-values and log fold changes as well as setting up the “for” loop to start at the first gene, continue to the second, and so on. By the end, I think at least some of the students who had been flummoxed came to understand how this code worked. I patted myself on the back.
With the lecture and tutorial out of the way, Atsu placed a call to Emmanuel to pick me up for a ride back to the guest house. He was so delighted to see that I was able to walk that he waived any fare for the ride! He’s a good guy. I spent the next few hours with my legs resting on the desk, higher than the rest of my body on the bed. I think I’ll be feeling more normal tomorrow!
October 19, 2018
We started with the hope for 30 students, but in the end, more than 60 students took part!
The final day of class came just in time, as my energy supply from being away from home for so long came to a close. When I arrived at the university, I found a magnificent pile of participation certificates to sign, courtesy of talented graphic designer Atsu Amewu (he also created the fabulous poster to publicize the Week of Bioinformatics event). I sat down at a desk to sign them all, and Samuel brought me the all-critical tall cup of tea.
Arriving in the lecture hall, however, brought on some new challenges; we had been compelled to shift to an adjoining hall because of a pre-scheduled event in our usual room. Unfortunately, the speaker system took a while to wake up, and the screen projector gave us fits! Eventually we connected an alternative VGA cable and powered up the screen, but the picture quality was definitely diminished because the VGA wire was not transmitting the red signal. With all the technical delays and another wait for certificate signing, we were delayed by around thirty minutes. (I should note that the technical challenges on this day continued as multiple power outages took place on campus, and the wifi network access faded into nonexistence as the day proceeded.)
I was pleased, though, with the students’ attention in the last lecture. The subject of biological pathways and networks is an interesting one, and they asked good questions in the course of the discussion. I added a couple silly stories along the way. I enjoyed talking about Trey Ideker, the genius behind the Cytoscape project. I noted that when you get to graduate school, you might be accustomed to being top dog, but you might well encounter somebody who has a much clearer vision of where they want to go than you do!
As we drew to a close, the head of the department, Prof. Peter Twumasi gave the students a rousing booster speech, declaring that bioinformatics is so important that he wants to see students regularly completing M.Sc. and Ph.D. degrees in Bioinformatics at KNUST Department of Biotechnology! I was delighted to see his enthusiasm kindling the students.
We mustered forty-nine faces on the final day of the workshop!
From there, we all assembled in a courtyard for the building, and a photographer captured an image of the whole group. When we came back inside, a great many of the students stopped for “selfies” with me. I might have been sleepy, but I found the energy to show my happiness. At first, I thought that I would have enough drive to go into Kumasi for a look at the palace, but ultimately I decided I needed to get off my feet. I retired to the KCCR guest house, and soon I was sound asleep.