Abstract: [krepp](https://doi.org/10.1101/2025.01.20.633730) is an alignment-free method for estimating distances of individual reads to many thousands of reference genomes in a scalable manner using k-mers (10x faster than bowtie2). Its distances are akin to alignment-based Hamming distances, and are extremely accurate. Using a statistical model, krepp can place genome-wide reads on large phylogenies (e.g., with >100,000 leaves) with errors lower than maximum likelihood placement (e.g., EPA-ng) of reads from marker genes. The software, together with a tutorial, is available on [GitHub](https://github.com/bo1929/krepp). In this talk, I will demonstrate how you can use krepp to index a set of reference genomes and then easily analyze metagenomic samples using this index.