Returning to Linköping

Things are wrapping up here ni London, we’ll be heading home in a few weeks. I have greatly enjoyed my stay here at Francis Crick, the Turner lab has been very friendly and is - simply put - amazing. I’m quite happy with how my analysis of the data turned out. First time working on scRNA-seq, which has been very fun. I could easily have poured another 3 months into the analysis but this will serve as a good starting point for whoever takes over. Overall I feel my own progress as a researcher has mainly been centered around 4 areas; scRNA-seq data analysis, coding proficiency, Rmarkdown and working on a HPC.

scRNA-seq data analysis

Aligning with Starsolo, loading count matrixes into Seurat and then try to mitigate empty droplets, doublets. It should be mentioned that I work with snRNA-seq as opposed to standard scRNA-seq. I have read lots of papers stating that 10X generally is quite tricky as it only captures a relatively small proportion of the transcriptome, giving rise to technical variation difficult to sort out from the true biological changes. This is definately something I see, and is made even more difficult we are not sure we’ll see any changes. Difficult but very fun!

Coding proficiency

I have people in my professional surroundings that are strong proponent of “if it works it works”. I agree with the sentiment, if a piece of code works then it its fine. However, in my case this lead to the code being full of repeating code instead of using functions or loops. This can of course be quite confusing and annoying to churn through, even if its well-annotated. Although I knew how to do functions/loops before, I have not really utilized them that much. This is an improvement I’m very happy with, I’ve felt its an area I need to improve on and I finally have. Its easy to push such improvements aside, given that the code still works fine.

Rmarkdown

For handing over the project I wanted to use Rmarkdown, as it seems like a simple way to output an understandable report but also including the code I have used. My understanding of Rmarkdown is still quite rudimentary but I feel I’m getting there. I remember when an old colleague asked me to use Rmarkdown to summarize an analysis I was working on. I remember feeling very lost and I’m sure the document was crap. I can probably find the document if I looked for it, but I’m not going to :D .

Working on a HPC

This was one of my worries coming here. What if I can’t figure out how to work on a high-performance cluster? It turned out to be very easy, as I have worked plenty on Ubuntu CLI before switching over to Thinlinc. At times it has been annoying though, queue times can be quite long (despite me being at work at 6 AM) and installing things is of course an issue. However, when I eventually got set up with the environment I wanted there were few hickups. Also, it is quite difficult to argue against being able to scale up to several TB of RAM.

Final note

So as a final note I’m very glad I had the opportunity to go here. I love London and the work environment here at the Crick, but landing in Linköping is going to feel very good.