Kari Lavikka

@karilavikka.fi

Researcher working on bioinformatics, WebGL-powered cancer genome visualization, and tumor evolution.

When doing copy-number analysis, have you ever struggled to interpret ASCAT sunrise plots and similar purity/p​​loidy grids? Have you wanted to compare how different solutions affect the allele-specific copy-number fit and better understand the fitting procedure?

I recently added a parameterizable arrow mark to GenomeSpy, my open-source toolkit for interactive genomic data visualization, and implemented a couple of fancy examples to showcase it. (1/3)

Parquet and Arrow are great columnar formats, but the abysmal performance of TextDecoder on Chrome (or maybe V8) completely kills any speed benefits when files contain lots of short, unique strings. Even #JavaScript CSV parsers are faster. It’s fine if you don’t need to access the string columns 🤷

It’s been a bit quiet on the GenomeSpy front, but I finally had time to implement interval selections and other fancy features. Vega-Lite–style conditional encodings are compiled into shader code, with selection tests running on the GPU—making interactions snappy. (1/2)

I've typically used R (with tidyverse) for analyses, but recently I've been familiarizing myself with Python. Pandas feels quite awkward to me, at least compared to dplyr. But @pola.rs is great! It works wonderfully with Parquet files too. ❤️

My first first-author paper is now published: doi.org/10.1093/giga.... GenomeSpy is a grammar-based visualization toolkit for genomic data, powered by efficient WebGL rendering. Check the paper for examples on how we use it to explore a large ovarian high-grade serous carcinoma dataset.

Deciphering cancer genomes with GenomeSpy: a grammar-based visualization toolkit

AbstractBackground. Visualization is an indispensable facet of genomic data analysis. Despite the abundance of specialized visualization tools, there remai

academic.oup.com