Fig. 1From: Lerna: transformer architectures for configuring error correction tools for short- and long-read genome sequencingWorkflow of Lerna. Lerna’s high level intuition is that by reducing the perplexity metric for the generated reads, we increase the alignment rate and the assembly quality (NG50). Our evaluation shows that a word-level Transformer LM, with a word length of 4 (|w| = 4) works best across all datasets and tools. Our algorithmic suite works on both NGS short reads and PacBio and Nanopore long readsBack to article page