Fig. 3

Overview of assembly algorithm. A Fragmented reads are overlaid and merged to reconstruct microbial genomes, with the longer reads enhancing the connections between each fragment. B MetaFlye, a LRS metagenome assembler, uses k-mer and repeat detection algorithm, which is particularly useful when detecting repeats inside a bubble. C Canu utilizes a form of k-mer algorithm known as adaptive MinHash k-mer weighting, in combination with an altered version of the greedy best overlap graph (BOG) algorithm. The greedy BOG algorithm, initially developed by Miller and colleagues, serves as the foundation for constructing the graph in Canu. Sequences with mutual best overlap are indicated with arrows going both ways. However, sequences G, H, and I have one-sided arrows, meaning that the best overlap regions are not mutual and are not included as part of the best overlap. D Hifiasm-meta utilizes k-mers to query reads and construct an overlap graph while retaining reads with rare k-mers, which typically correspond to low abundance sequences