In the separate experiments, the GF and RC methods detected similar dietary treatment effects on methane emission (expressed as g/d and g/kg of DMI), although the magnitude of the differences varied between experiments.
Interestingly, both alignment methods detected similar twelve conserved motifs (Figure S1 and Figure S2 in Supporting file S1).
Both methods detected similar trends of expression (up or down regulation) although the fold changes may not be the same.
We found that both methods detected similar patterns of expression for the 17 upregulated and three downregulated genes.
For 40 out of 50 genes tested, the two methods detected similar expression trends across development (Pearson r ≥ 0.70; Additional file 2).
The one-step BLMM and two-step mLASSO methods detected similar sets of QTL given that a large number of individuals were studied and the adaptive shrinkage method of BLMM appeared to allow reasonably large effect estimates in case QTL were significant.
In contrast to the observations made on the A-set, BLMM did not even detect a single suggestive QTL in the statistically low-powered SNP+AFLP dataset, whereas the mLASSO method detected similar numbers of suggestive or significant QTL in both the S+A-sets and A-sets (Table 3).
Regarding local structure comparison methods to detect similar active sites, several 'template-based' methods have been reported (Barker and Thornton, 2003; Chou and Cai, 2004; Fetrow and Skolnick, 1998; Ivanisenko et al., 2004; Kleywegt, 1999; Laskowski et al., 2005; Stark and Russell, 2003; Torrance et al., 2005; Wallace et al., 1997).
The papers of Janes et al [ 7- 9] are close to our aims as the implementation of co-clustering methods to detect similar expression patterns, e.g., [ 10].
