- Consensus sequence using Racon
- Consensus sequence using minipolish
- Consensus sequence using Medaka
- Error correction using Pilon
Despite recent advances one of the main problems with Long-Read technologies is the increased error rate, especially in Oxford Nanopore data. One way to address this is the building of a consensus sequence. To build a consensus sequence all reads are aligned to the assembly and for each position in the assembly the nucleotide with the highest number of reads supporting it is chosen as the consensus nucleotide.
Additionally, some tools try to further improve the consensus sequence, e.g., through resolution of insertion-deletion errors (INDELS) or by taking into account the raw signal instead of just the basecalled reads.
Often higher accuracy can be achieved by repeating the consensus-building several times and/or combining several consensus and error correction tools.