Statistical analysis of association between virulence-associated phenotypic traits and in vivo virulence.

aTraits ordered by proximity for better comprehension. Data used in the case of strains isolated from dietary supplements are shown in Table 5. Phenotypic traits of control strains, simplified following the codes in Table 5, are: YJM128 (+, +, +, +, –, +, +, n.d., n.d., n.d., n.d.), W303 (+, –, –, +, –, –, –, n.d., n.d., n.d., –), CECT 10.431 (+, –, –, –, –, –, –, n.d., n.d., n.d., n.d.). Data of YJM128 have been obtained previously [28]. “+” and “–” were coded numerically as “1” and “0” respectively for statistical analysis.bIn vivo virulence determined by intravenous inoculation of strains D2, D4, D14, D23, YJM128, W303 and CECT 10.431 in BALB/c mice. Virulence was simplified as follows: high burdens levels in brain (3.18 to 5.31 mean log10 CFU) at 7 and 15 days post-infection were considered + and lower burdens levels and the inability to colonize was considered –.cA proximity value of “1” indicates the maximum association and value “0” indicates the opposite. Jaccard proximities were calculated excluding joint absences from consideration.dValues of “1” indicates perfect correlation and “0″ indicates no relationship.n.c. Spearman’s correlation cannot be calculate because this trait shows constant values (see Table 5).