Figure 2.

Summary of probabilistic sequences construction. Nucleotide-level probabilistic sequences can be generated from a single FASTQ or SAM file using the sequencing quality information (left). In the case of multiple FASTQ or SAM the user can average the sequencing quality information beforehand (center). When multiple FASTA files are available, the probabilities can be directly informed from the frequencies of nucleotides at each position (center). In the case of a single FASTA file or consensus FASTQ file, the user can assume a probability model for the distribution of sequencing errors (right). Sequence-level probabilistic sequences may be obtained from the nucleotide-level ones, for example by selecting the n most probable sequences (bottom).