Consensus modeling

The DNA sequence analysis of specific genes has allowed the recognition of a number of sequences important in gene transcription. From the large number of bacterial genes smdied it is possible to construct consensus models of transcription initiation and termination signals. 

In 2005, a comprehensive comparison of LCIA toxicity characterization models was initiated by the United Nations Environment Program (UNEP) and the Society for Environmental Toxicology and Chemistry (SETAC) in their life cycle initiative. The main objectives of this effort were to [9] (1) identify specific sources of differences between the models results and structure (2) detect the indispensable model components and (3) build a scientific consensus model from them, representing recommended practice. 

This procedure assessed whether some of the different descriptors used by different equations were intercorrelated and, therefore, interchangeable [59]. The remaining diverse QSAR equations were further classified by size (number of descriptors they include). The best equations of each encountered size were kept for final validation with the VS molecules and for further analysis. Consensus models featuring average predictions over these equations were also generated and validated. We focus here on the discussion of the minimalist overlay-independent and overlay-based QSAR models, each including only six descriptors, and refer to the optimal consensus model of the overlay-based QSAR approach families for comparative purposes. 

There are two possible application strategies for the use of 4D-QSAR models as a VHTS. The first is to take a collection of (manifold) 4D-QSAR models and create a consensus 4D-QSAR model. The consensus model is evaluated for each molecule using all of the individual 4D-QSAR models ... 

This strategy was successfully applied in QSAR [62] and preliminary results have demonstrated an increased accuracy in log Poet prediction when consensus models were derived by neural network using as input eight well known prediction values [63]. 

Zuaboni, D., Cleva, C. and Carrupt, P.A. (2009) Consensus models and metamodels for the prediction of log P using neural networks. Private communication. 

E-state indices, counts of atoms determined for E-state atom types, and fragment (SMF) descriptors. Individual structure-complexation property models obtained with nonlinear methods demonstrated a significantly better performance than the models built using MLR. However, the consensus models calculated by averaging several MLR models display a prediction performance as good as the most efficient nonlinear techniques. The use of SMF descriptors and E-state counts provided similar results, whereas E-state indices led to less significant models. For the best models, the RMSE of the log A- predictions is 1.3-1.6 for Ag+and 1.5-1.8 for Eu3+. 

Yongye AB, Byler K, Santos R et al (2011) Consensus models of activity landscapes with multiple chemical, conformer, and property representations. J Chem Inf Model 51(6) 1259-1270... 

Extensive investigations on the catalytic mechanism of classical peroxidases resulted in a consensus model involving five different iron species [30, 31], These species are ferrous, ferric, Compound I, Compound II, and Compound III (Fig. 11.1). As discussed in Chap. 5, after the reaction of ground state (GS) Femporphyrin with H202, Compound I (Cl) is formed, a cationic oxob e,vpor-phyrin-based Ji-free radical. Electron paramagnetic resonance (EPR) studies established that, in peroxidases of classes I and III, the second oxidation equivalent in Cl is present as a porphyrin-based free radical [32, 33]. In peroxidases from fungal sources, electron abstraction from the protein results in the formation of a different species with the free radical based in a residue close to the porphyrin. 

The Decision Forest (DF) is a DT-based consensus modeling method that was conceived as a method for combining heterogeneous yet comparable trees that fully captures the association between molecular structure and biological activity (Figure 6.4) [59], The heterogeneity requirement assures that each tree uniquely contributes to the combined prediction whereas the quality comparability requirement assures that each tree equally contributes to the combined prediction. 

Gramatica P, Pilutti P, Papa E. Validated QSAR prediction of OH tropospheric degradation of VOCs Splitting into training-test sets and consensus modeling. I Chem Inf Comput Sci 2004 44 1794-802. 

---