Here is a brief description about the strategy used for reconstructing multi-gene superclusters of *B. cereus* group isolates.

Data from the 5 MLST schemes were used to construct superclusters of genetically closely related strains using a tree-independent procedure. Isolates were grouped if they shared identical sequences at all loci of a given MLST scheme (i.e., belong to the same sequence type (ST) using 6 or 7 loci, depending on the scheme). The grouping was done in a recursive manner. That is, if strain A had identical sequences to strain B based on scheme 1, and was identical to strain C based on scheme 2, then strains A, B, and C were grouped. All strains sharing an ST with B and all strains sharing an ST with C were subsequently added to the supercluster, then all strains identical to the latter strains were added, and so on until no further isolate sharing an identical ST with the strains currently in the group could be found. As a result of this procedure, every isolate of a given supercluster shares identical MLST data to at least one other isolate in the group. Note also that all superclusters that were built by this procedure are mutually exclusive, i.e., a given isolate belongs to one and only one group. The algorithm described above corresponds to solving connected graphs in mathematics.

Below is an example of a reconstructed multi-scheme MLST supercluster. Strains sharing identical MLST profiles are connected by lines colored by scheme. While strains B.c_MM_2, B.c_BFDA96145, and B.c_ATCC10987 all have identical profiles for the Priest (P) scheme (green lines), B.c_B309 and B.c_B06_009 are identical to B.c_ATCC10987 based on the Helgason (H) and Tourasse-Helgason (TH) scheme, respectively (red and blue lines).