Top 10 European Biopharma Clusters
As Brexit Casts Uncertain Shadow, Continent’s Industry Continues Growth Trajectory
Diet, the Gut Microbiome, and Colorectal Cancer: Are They Linked?
Updates on Microbial Associations and Therapeutic Implications
Precision CRISPR Tools Make the Cut
Genome-Editing Tools Modify More than Genomes
Non-Viral, Plasmid-Free CAR T-Manufacturing Platform Begins Preclinicals
Goal Is the Development of Faster, More Affordable Therapies
How to Use t-SNE Effectively
Especially within the context of high-dimensional single-cell RNA-sequencing datasets, biological research articles increasingly include “t-SNE plots”—plots designed to convert multidimensional datasets into lower-dimensional representations based on similarities in the data. For example, data points corresponding to different “cell types” are often depicted as distinct clusters in a 2D plot. While these plots can be very useful, they are accompanied by many caveats that are often unknown to the casual observer. Enter the article, “How to Use t-SNE Effectively” on Distill.pub (a website that provides articles to simply explain concepts in machine learning). This article uses interactive simulations to teach site visitors the effects of altering different parameters (e.g., points per cluster, dimensions, and “perplexity”) used by the t-SNE clustering algorithm. A nice text explanation and embedded illustrations drive home the main points.