Blog

In cross-species single-cell workflows, we often transfer cell type annotations from a well-characterized reference (e.g., human) onto a query species (e.g., mouse). The implicit assumption is that every query cell has a counterpart somewhere in the reference. But what happens when it doesn’t? [Read More]

Like many researchers, I work across fast hot storage (scratch or SSD) and slow, cost-effective cold archives. Managing data across these tiers often means moving files between large, tape-backed archival systems and higher-performance disks for active analysis. My institution uses Globus for these transfers, which generally offers a smooth experience,... [Read More]

While working on downsampling of a single-cell (sc) dataset i.e., performing binomial thinning on raw UMI counts across matrices of 17,000 genes by 150,000 cells (2.55 billion entries, ~127 million nonzero elements at 5% sparsity), I expected that updating the data in place, chunk by chunk, would be straightforward and... [Read More]

When analyzing dozens or thousands of omics datasets, it is essential to parallelize the compute. In this post, I showcase a couple of ways for large-scale batch analysis on high-performance computing (HPC), specifically focusing on cases where each sample can be processed independently using the same workflow and comparable computational... [Read More]

In this post, I will highlight some of the ways in which you can manipulate the information displayed as part of the command prompt. The default shell in all Apple computers and laptops changed to zsh from macOS Catalina and onwards. Zsh is a specific type of shell just as... [Read More]

This post will walk you through all the steps needed to setup your own website using Jekyll and Github pages documenting set up a local version of the blog repository (repo) in a virtual environment using conda. [Read More]