Fig. 3From: An efficient error correction algorithm using FM-indexComparison of traditional seed-and-extend alignment versus compressed alignment using FM-index extension. a Without compression, all reads containing the seed will be individually aligned against the query in order to compute overlap and similarity; b With compression, all reads containing the seed will be represented by different paths during FM-index extension. These reads will be gradually extended (extracted) from the seed sequence at both directions via FM-index extension, while identical sequences are still compressed in a single path. Simultaneously, each newly extended base will be aligned against the query readBack to article page