Fig. 1

Schematic overview of LoMA's algorithm. A. The main process to compose a consensus sequence (CS) is shown. Reads are laid out based on all-to-all alignment using minimap2 [16]. The layout is then subdivided into small divisions for the successive procedure to determine partial CSs based on multiple alignments using MAFFT for each division [17]. The entire CS is obtained by concatenating these small divisions. B. The read classification process is shown. Reads constituting a CS are aligned to the CS using minimap2 [16]. LoMA predicts heterozygous loci in the region based on the number of reads supporting a SV, and the reads derived from each estimated haplotype are gathered (Type I and II read sets). Both sets return to the main process (A) after this step