Fig. 2
From: metaMIC: reference-free misassembly identification and correction of de novo metagenomic assemblies
metaMIC outperforms ALE and DeepMAsED in identifying misassembled contigs in simulated metagenomic datasets. a–d The performance of the three tools on the CAMI-medium (M) and high-complexity (H) communities (a), CAMI2-Skin (b), CAMI2-Gut (c), and simulated virome dataset (Sim-Virome) (d). e The AUPRC scores of metaMIC on test datasets assembled by MEGAHIT or IDBA_UD (Test assembler), where metaMIC were trained on contigs from training datasets assembled by MEGAHIT, IDBA_UD, or jointly by MEGAHIT and IDBA_UD (MEGAHIT+IDBA_UD)