Posts

While having access to HPC, resource is not infinite, and our HPC reasonably manage job queues by how much computation you have requested recently. This provides incentive to only request enough for whatever task you work on, but for me, it hasn’t been easy: Estimating resource usage is not trivial, and trying to be as frugal as possible would sometimes result in time-consuming jobs that are terminated after several hours or days due to insufficient memory or running time requested.

If you updated Seurat recently, you might find your FeaturePlot() and DimPlot() giving plots that look pixelated instead of the circular dots we are used to. Don’t panic, this is because in a recent update, raster = TRUE becomes default for FeaturePlot() and DimPlot() when your dataset contains more than 50,000 cells.

I consider myself to be moderately addicted to coffee: I slurp a mug of coffee to start my day, and when in certain days, I take another boosting dose in the afternoon and see my hands trembling.

The problem When compiling Emacs 28.05, make bootstrap failed with fatal error: 'libxml/tree.h' file not found while autogen.sh and configure were successfully executed. Troubleshooting The error message pointed out the culprit, libxml.

I’ve been working with pySCENIC in the recent couples of months. It’s a tool that is meant to uncover gene sets that are controlled by transcription factors and suggests the underlying logic of how genes are regulated.

I used to dislike dependency management because there are so many things that require root or are mysteriously incompatible with each other. However, I recently came to realize that the existence of dependency itself is a delicate and elegant system despite being messy at times.

Self-handicapping can be quite productive on those non-urgent things. Recently, I noticed a LaTeX template, Awesome CV, so I was thinking trying it to replace Modern CV, which has accompanied me for around three years.

I am exaggerating, but sometimes stringsAsFactors is almost this deadly. I work with genomic data, and a common quest in my job is to identify interesting features (in most cases, genes) from a pool of 25,000+.

After struggling for a while, I decided to move from Medium and switch to blogdown. While Medium is a beautiful platform for blogging, its philosophy seems to fit less well when there are more than articles to host.

The more I work with Linux, the more I encounter dependency issues. This is of course not too big a surprise, but it can be painful especially when you aren’t sudo, so the most obvious solution does not work for you.