Assembly combines overlapping sequencing reads into longer sequences called contigs. It offers a different view of a community from read-level profiling: contigs can preserve longer stretches of genomic context, and the quality of their assembly is evaluated on the basis of different criteria, as it may depend on sequencing depth, community complexity, and the sequencing technology used. We will introduce the foundations of short-read and long-read assembly and discuss the trade-offs that shape the resulting contigs.We will then tour through the MGnify’s assembly pipeline and methods for the quality assessment of the results. In addition, we will explore how sequence features are annotated, including taxonomic and functional information, and use the contig viewer visualisation on the MGnify website to inspect annotations in their genomic neighbourhood. A worked example will show how a biological question can guide exploration of an assembly, while keeping in mind that individual contigs may be incomplete and do not necessarily represent a whole genome.This event is part of the webinar series “Making sense of metagenomics data with MGnify”. You can follow the link for more information about the series and its webinars.