Fig. 2
From: Multi-CSAR: a multiple reference-based contig scaffolder using algebraic rearrangements
Schematic workflow of Multi-CSAR: a A target genome T={c1,c2,c3,c4} and three single reference-derived scaffolds S1=(+c1,+c2,+c3), S2=(+c2,+c3,+c4) and S3=(−c2,−c1,−c4,−c3) that are assumed to be obtained by applying CSAR on three reference genomes R1,R2 and R3, respectively, with equal weight of one. b The contig adjacency graph G constructed by using S1,S2 and S3, where the dashed lines denote the edges with zero weight. c A maximum weighted perfect matching \(M=\left \{\left (c_{1}^{h}, c_{2}^{t}\right), \left (c_{2}^{h}, c_{3}^{t}\right), \left (c_{3}^{h}, c_{4}^{t}\right), \left (c_{4}^{h}, c_{1}^{t}\right)\right \}\) derived by applying Blossom V on G. d By removing the minimum weighted edge \(\left (c_{4}^{h}, c_{1}^{t}\right)\) from M, we obtain \(M^{\prime }=\{(c_{1}^{h}, c_{2}^{t}), (c_{2}^{h}, c_{3}^{t}), (c_{3}^{h}, c_{4}^{t})\}\) such that M′∪C contains no cycles, where the dotted lines denote the edges in C. e The final scaffold (+c1,+c2,+c3,+c4) of T constructed based on the edge connections in M′