Taye Temitope Alawode and Kehinde Olukunmi Alawode / Elixir Network Engg. 110 (2017) 48206-48212 48206 Introduction A major challenge faced by natural products chemists in the process of drug development is elucidation of the structure of isolated compounds. Sesquiterpenes are formed from countless biogenetic pathways and therefore produce several types of carbon skeletons [1,2]. This makes elucidation of their structures very challenging. Eudesmane-type compounds are one of the most representative skeletons of sesquiterpenes. Emerenciano and co-workers have developed and applied the expert system, SISTEMAT (based on artificial intelligence programs) in the elucidation of structures of several classes of compounds including sesquiterpenes [3,4], lactone sesquiterpenes[5], diterpenes [6] and triterpenes [7]. Oliveira and co workers [4] demonstrated the use of SISTEMAT in obtaining useful rules of 13 C spectral analysis and its use as an auxiliary tool in the process of structure elucidation for eudesmanes. The same work presented a review on the 13 C NMR data of eudesmanes. Part of the 13 C NMR chemical shift data used in our previous studies and the current one are obtained from this publication. The structure of any natural product is conventionally divisible into three sub-units: the skeletal atoms, heteroatoms directly bonded to the skeletal atoms or unsaturations between them and secondary carbon atoms, usually bonded to a skeletal atom through an ester or other linkages [5]. The skeletal structure common to all Eudesmane compounds is shown in Fig.1. Fig 1. Eudesmane skeleton. Typical substituents found in Eudesmane compounds are presented in Fig. 2. Our Contribution In a previous study [8], we have shown that when Generalized Regression Neural Network (GRNN), is trained using the 13 C chemical shift values for each of the 15 positions on the eudesmane skeleton as input and the various possible substituents as the target, GRNN could identify the substituents in each position on the Eudesmane skeleton of unknown compounds (when the 13 C values for each position on the Eudesmane skeleton of the each unknown eudesmane is supplied to the system). We have also applied scatter plot as a tool to determine the 13 C chemical shift ranges (for each of the 15 carbon positions on the Eudesmane skeleton) over which different substituent types may exist allowing the determination all the possible structures consistent with a particular set of spectroscopic data [9]. However, full elucidation of structures of unknown compounds using the described procedures could not be carried out since the studies were based on the premise that the 13 C skeletal data of the compounds whose substituents were being determined are known. In the present work, we predict the 13 C chemical shift values on the skeleton (C 1 -C 15 ) of novel eudesmane skeleton using GRNN. We thereafter proceeded to utilize the predicted skeletal values to determine the substituent types in each position on the eudesmane skeleton employing the principles demonstrated in the previous works. The degree of accuracy of GRNN and scatter plots in determining the substituents on each position of the skeleton of the test compounds were compared. In utilizing the GRNN and scatter plots in predicting the substituent types, the original data set utilized in our previous publications were expanded in order to accommodate new substituent types encountered in the compounds employed in skeletal data prediction. Tele: +234-7056712384 E-mail address: onatop2003@yahoo.com © 2017 Elixir All rights reserved ARTICLE INFO Article history: Received: 11 July 2017; Received in revised form: 1 September 2017; Accepted: 11 September 2017; Keywords GRNN, Scatter plots, Structural elucidation, 13 C, Eudesmane sesquiterpenes. Structural Elucidation of Eudesmane Sesquiterpenes using GRNN and Scatter Plots Taye Temitope Alawode 1 and Kehinde Olukunmi Alawode 2 1 Department of Chemical Sciences, Federal University Otuoke, Bayelsa State, Nigeria. 2 Department of Electrical and Electronic Engineering, Osun State University, Osun State, Nigeria. ABSTRACT This study seeks to achieve a complete elucidation of structures of unknown Eudesmane sesquiterpenes from their 13 C values. The 13 C values for each of the fifteen (15) positions of the skeletons of the Eudesmane compounds were predicted using Generalized Regression Neural Network (GRNN). From the predicted 13 C values, GRNN and Scatter Plot methods were used to predict the substituents attached to each position on the skeleton of the Eudesmane compounds. Recognition of the test compounds ranged between 40 and 100%. GRNN and Scatter plots demonstrated great potential for use in the structural elucidation of unknown compounds from 13 C values. © 2017 Elixir All rights reserved. Elixir Network Engg. 110 (2017) 48206-48212 Network Engineering Available online at www.elixirpublishers.com (Elixir International Journal)