ALL-QSAR

ALL-QSAR automated lazy learning quantitative structure-activity relationships... 

Zhang, S., Golbraikh, A., Oloff, S., Kohn, H., Tropsha, A. A novel automated lazy learning QSAR (ALL-QSAR) approach method development, applications, and virmal screening of chemical databases using validated ALL-QSAR models. 

A widely used approach to establish model robustness is the randomization of response [25] (i.e., in our case of activities). It consists of repeating the calculation procedure with randomized activities and subsequent probability assessments of the resultant statistics. Frequently, it is used along with the cross validation. Sometimes, models based on the randomized data have high q values, which can be explained by a chance correlation or structural redundancy [26]. If all QSAR models obtained in the Y-randomization test have relatively high values for both and LOO (f, it implies that an acceptable QSAR model cannot be obtained for the given dataset by the current modeling method. 

Any QSAR method can be generally defined as an application of mathematical and statistical methods to the problem of finding empirical relationships (QSAR models) of the form ,- = k(D, D2,..., D ), where ,- are biological activities (or other properties of interest) of molecules, D, P>2,- ,Dn are calculated (or, sometimes, experimentally measured) structural properties (molecular descriptors) of compounds, and k is some empirically established mathematical transformation that should be applied to descriptors to calculate the property values for all molecules (Fig. 6.1). The goal of QSAR modeling is to establish a trend in the descriptor values, which parallels the trend in biological activity. In essence, all QSAR approaches imply, directly or indi-... 

To validate the equilibrated receptor, its potency to predict free energies of binding (AGpiej) is examined. Therefore, classical QSAR methods such as cross-validation via leave-X%-out analyses and/or prediction of activity for an external set of compounds (test set) are accomplished. Since all QSAR models are typically constructed to predict properties of new or even virtual molecules, model validation with an external test set reflects reality best (unbiased or biased random selection of training set and test set ligands is supported by the software). 

A subtle distinction can be made between QSARs and the PMs associated with physicochemical tests. The distinction is that while any PM (associated with a physicochemical test) could also be called a QSAR, not all QSARs could also be called PMs. For example, QSARs can also be based on theoretical descriptors (e.g., topological indices) or on experimental properties that are themselves more easily predicted than measured (e.g., the octanol-water partition coefficient). Furthermore, QSARs developed for the prediction of physicochemical and in vitro end points would not be regarded as PMs. 

All QSARs selected for validation should meet a minimum set of criteria, which could be called QSAR development criteria ... 

All QSARs should be associated with a clear scientific and regulatory purpose. A possible outcome of the peer-review step is the recommendation that the validated QSAR should be considered for inclusion into a regulatory framework. In such a case, the recommendation, and all of the supporting evidence, should be forwarded to the appropriate regulatory body (or bodies) for consideration. In the EU, if the QSAR is considered suitable for the assessment of chemicals, deliberations take place at a technical level by the National Coordinators for Testing Methods, and subsequently, a decision is taken at the policy level by representatives of the EU Competent Authorities for Directive 67/548/EEC. 

VIII) As shown in Sect. 5.4 the log Kqw value of ionizable organic chemicals depends on the pH. Therefore, the BCFl values of these chemicals also depend on the pH. Consequently, the BCFl data of these ionizable organic compounds have to be correlated with their log Kqw values at the pH (normally about 7) of the water, which prevailed during the bioconcentration test. In most, if not all QSARs this fact was so far not considered. 

All QSAR analysis was carried out using the program ALMOND 3.2.0 and grid alignment independent descriptors (GRIND). 

Most, if not all, QSAR methods require selection of relevant or informative descriptors before modeling is actually performed. This is necessary because the method could otherwise be more susceptible to the effects of noise. The a priori selection of descriptors, however, carries with it the additional risk of selection bias [73], when the descriptors are selected before the dataset is divided into the training and test sets (Figure 6.6A). Because of selection bias, both external validation and cross validation could significantly overstate pre-... 

When working with pseudo-receptors, and in general with quantitative structure-activity relationships (QSAR) of any dimension, a word of caution is necessary with respect to the biological data that is used. These should preferably constitute binding affinities from a single laboratory, a prerequisite which is also true for all QSAR studies. Since the receptor models simulate interaction events (AH) in a highly simplihed manner, the experimental data which are combined with them in a correlation analysis mnst be as close to the molecular level as possible. It is therefore nonsense to correlate the calculated interaction energies... 

Equations (9) and (10) constitute the fundament of all QSAR studies. Since 1964, they have remained essentially unchanged, with the exception of two minor modifications. Improvements resulted from the combination of Hansch equations with indicator variables [22], which may be considered as a mixed Hansch/Free-Wilson model (Eq. (11)) [23], and from the formulation of a theoretically derived nonlinear model for transport and distribution of drugs in a biological system, the bilinear model (Sec. 4 Eq. (30)) [24] ... 

Strictly speaking, still today there is no way to apply eq. 1 to biological data. All QSAR equations correspond to eq. 2, because only the differences in biological activities are quantitatively correlated with changes in lipophilicity and/or other physicochemical properties of the compounds under investigation. 

In all QSAR equations reported here, n is the number of data points, r is the correlation coefficient, s is the standard deviation, q is Cramers coefficient to account for the variance in the activity [69] and the data within the parentheses are 95% confidence intervals. F is the F-ratio between the variances of observed and calculated activities. 

All QSAR techniques assume that (1) all the compounds being studied bind to the same biological target noncovalently (2) structurally similar compounds are similarly oriented at that common receptor site and (3) the dynamics of the system can be neglected. The methods differ, however, in the way they describe the compounds and in how they detect the relationships between 3D properties and bioactivity. As with all QSAR methods, 3D-QSARs are often used to predict the potency of compounds not yet tested. 

The underlying concept of all QSAR analyses is the additivity of substituent group contributions to biological activity values in the logarithmic scale. This additivity comes from the fact that QSAR models are linear free-energy related. All... 

The Tsar interface provides a complete system for QSAR analysis. By including the full molecular description in a spreadsheet together with the computed and measured data, all QSAR calculations are made possible. Thus, the program can integrate property calculations and data analyses in a single environment, providing simple access to these methods, even for novice users. 

Finally, all QSAR, Spectral-SAR and Qua-SAR computational results may be collected and resumed by associate spectral scheme for evolution of the fittest molecular structures along the endpoint models for the (M=)3 selected mechanistic paths of actions, see Figure 3.25. Note that algebraic correlation environment was chose as the vertical indicator... 

---