The phylogenetic tree was constructed using the UPGMA clustering method (MEGA 4).
The revision of the original clustering pipeline used to build FCPD [ 22] was motivated by a few factors, including the presence of many mega clusters with over 100 members, singlet clusters, and clusters that did not match families in Nelson's P450 databases.
Around an Euclidian distance of 0.8, one of the mega clusters fell into 6 clusters, and the other into 3.
Not only is the overall number of disasters increasing, each calendar year seems to be punctuated by a cluster of mega-disasters like this year's record-breaking fires in California or last year's record-breaking Atlantic hurricane season, which saw not one, but three massive hurricanes sowing nothing but havoc in their path.
Additionally, the original FCPD classification generated several mega clusters, underscoring the need for optimizing clustering parameters.
Only ten accessions formed three additional side groups to the two main mega clusters (see Additional file 4).
Most of the remaining accessions fell into two mega clusters, similar to the dendrogram for phenotypic data.
Diversity analysis on both the phenotypic and the molecular level suggested that the core collection is distributed over two mega clusters.
Visual inspection of the dendrograms showed that while the grouping of the accessions to the mega clusters was consistent between the dendrograms for phenotypic and genotypic similarity, the arrangement of the entries in the subgroups was highly different.
Based on the OTU-distribution the Pearson distance correlations were calculated and resulting distance matrices used for cluster analysis with MEGA 5. Sequence information was also used for phylogenetic tree reconstruction (UPGMA or neighbor-joining) based on MEGA 5 and interpretation of phylogenetic trees (Cluster analysis and principal coordinate analysis) performed with FastUniFrac [ 44].
When the mismatch sequence was included in MEGA cluster analysis, it separated from all other (wild and cultivated) alleles with 87% bootstrap support, and the mean D among sequences increased 8-fold (D = 1.5 for n = 10 alleles versus D = 12 for n = 11 alleles).
