Superposing Significant Interaction Rules (SSIR) Method: A simple Procedure for Rapid Ranking of Congeneric Compounds

Abstract

The Superposing Significant Interaction Rules (SSIR) method is revised and implemented. The method is a simple combinatorial procedure, which deals with in situ generated rules among a dichotomized congeneric molecular family, selecting the most probabilistically relevant ones. The mere counting of the number of relevant rules attached to new compounds generates a molecular ranking useful for database filtering, refinement and prediction. The algorithm only needs for a symbolic molecular representation and this allows for mining the database in a confidential manner. Third parties will not know the real compounds that are on the way to be worked out. The procedure is tested for a complete series of substituted amino acids. Areas under the receiver operating characteristic (AU-ROC) are always greater than 0.9 for all the following tried protocols: training, leave-one out, balanced leave-two-out and 5-fold cross validations and, finally, a stochastic series of calculations combined with a randomization test.

Investigation of Volatiles in Cork Samples Using Chromatographic Data and the Superposing Significant Interaction Rules (SSIR) Chemometric Tool

“…of the HIV-1 reverse transcriptase [33] and a family of substituted peptides [34]. The procedure is related to the design of experiments field, and has been applied to obtain fast qualitative preliminary results before the development of a full design [35].…”

Ranking Series of Cancer-Related Gene Expression Data by Means of the Superposing Significant Interaction Rules Method

“…The generation of SSIR rules is carried out in a systematic and combinatorial mode (despite non exhaustive procedures being also envisaged). In References [27,29], the reader can find additional details of the algorithm and its rationale. An SSIR rule consists of the specification (by inclusion or exclusion) of some levels for sites.…”

“…Eventually, a third one generates 2 k binary numbers, constituting a flag for encoding of all possible combinations of negations or non-negations. See Reference [29] for exhaustive information on how to implement the appropriate algorithms.…”

“…The procedure just mentioned constitutes a simple fitting or training method (together with an eventual external testing). Distinct test or validation schemes can be easily implemented in SSIR [27,29]. Here, leave-one-out (L1O) or randomization test results will be presented below.…”

“…At the end of each calculation, the left-out case is evaluated. However, this procedure does not scale directly according to the above 30-min timing because, when performing the training, the information related to the rules are kept in memory or a disk for later use [27][28][29]. In doing so, it is quite fast to simulate inclusions or exclusions of individuals when the rules need to be reevaluated during the L1O protocol.…”