Above and beyond the variability included in the base model, blood pressure and obesity-related variables explained most of the variability in BrachD (change in adjusted R of 13 and 11%, respectively, from base model) and PWV (change in adjusted R of 7 and 4%, respectively, from base model).
Parent material and some climatic indexes were the most influential variables explaining the scaling properties of particle size distributions.
Multivariate analysis confirmed the outdoor terrace area, the season, the type of enclosure, and the number of smokers as the most relevant variables explaining nicotine concentration (R2=0.396).
We show that while elevation and geology are the most important variables explaining the spatial distribution patterns of vegetation for both models, the influence of the climatic and other variables on the vegetation at the two levels differ.
Overall, redundancy analysis (RDA) indicated that soil surface area, percent deciduous canopy cover, elevation, as well as soil albite content were the most significant variables explaining together 56% of the total variation in BC weathering rates.
The depth and exposition are the most important variables explaining the coastal temperature gradients from the innermost to the outermost areas in springtime when water is heated by increasing solar radiation.
Multivariate analyses of Δ Dq versus all variables (Alu, G+C, CGI, LINE, MIR, MER, LTRs, nCoding, nNonCoding, exons, genes, and SNPs) per chromosome were carried out and for each case the most relevant variables explaining Δ Dq were selected.
Multivariate analysis revealed that the most important variables explaining tobacco use among study participants were the use of khat (OR = 14.05; P = 0.000), smoking status of friends (OR = 2.25; P = 0.000), and substance use by friends (OR = 1.81; P = 0.001).
In the same way, we analyzed Δ Dq for the whole genome, obtaining again (G+C), Alu, CGI as the most relevant variables explaining Δ Dq (Additional file 4).
The results for the model of all students suggest that the most important variables explaining tobacco use among study participants were student's khat use status (OR = 14.05; P = 0.000), friends' smoking status (OR = 2.25; P = 0.000), and friends' substance use status (OR = 1.81; P = 0.001).
