Genetic Algorithm-partial Least Squares Research Articles

Rapid and Nondestructive Quantification of Trimethylamine by FT-NIR Coupled with Chemometric Techniques

This paper focused on the quick and nondestructive evaluation of trimethylamine (TMA-N) in fish storage which is sequent to its freshness, the key for controlling the quality and safety of fish products by combining Fourier transform near-infrared (FT-NIR) and chemometric techniques. Calibration models of fish freshness were established using three multivariate chemometric methods—partial least square (PLS), synergy interval PLS (Si-PLS), and genetic algorithm PLS (GA-PLS) for quantitative prediction of TMA-N in fish. Results of the developed model were estimated using the correlation coefficients of the prediction (Rp) and calibration (Rc); root mean square error of prediction (RMSEP) and the ratio of sample standard deviation to RMSEP (RPD). The established model’s performance achieved 0.943 ≤ Rp ≤ 0.977 and 4.25 ≤ RPD ≤ 4.30. The model’s prediction strength improved in the order PLS < Si-PLS < GA-PLS. GA-PLS significantly improved the prediction of TMA-N prediction with RMSEC = 5.08 and Rc = 98.28 for the calibration data whereas the prediction set gave an RMSEP = 5.10 and Rp = 97.70. FT-NIR spectroscopy combined with GA-PLS technique may be employed for rapid and non-invasive quantification of TMA-N in fish for monitoring safety and quality.

Rapid Determination of Chlorogenic Acid, Luteoloside and 3,5-O-dicaffeoylquinic Acid in Chrysanthemum Using Near-Infrared Spectroscopy.

The feasibility of near-infrared spectroscopy (NIR) to detect chlorogenic acid, luteoloside and 3,5-O-dicaffeoylquinic acid in Chrysanthemum was investigated. An NIR spectroradiometer was applied for data acquisition. The reference values of chlorogenic acid, luteoloside, and 3,5-O-dicaffeoylquinic acid of the samples were determined by high-performance liquid chromatography (HPLC) and were used for model calibration. The results of six preprocessing methods were compared. To reduce input variables and collinearity problems, three methods for variable selection were compared, including successive projections algorithm (SPA), genetic algorithm-partial least squares regression (GA-PLS), and competitive adaptive reweighted sampling (CARS). The selected variables were employed as the inputs of partial least square (PLS), back propagation-artificial neural networks (BP-ANN), and extreme learning machine (ELM) models. The best performance was achieved by BP-ANN models based on variables selected by CARS for all three chemical constituents. The values of rp2 (correlation coefficient of prediction) were 0.924, 0.927, 0.933, the values of RMSEP were 0.033, 0.018, 0.064 and the values of RPD were 3.667, 3.667, 2.891 for chlorogenic acid, luteoloside, and 3,5-O-dicaffeoylquinic acid, respectively. The results indicated that NIR spectroscopy combined with variables selection and multivariate calibration methods could be considered as a useful tool for rapid determination of chlorogenic acid, luteoloside, and 3,5-O-dicaffeoylquinic acid in Chrysanthemum.

NIR Spectroscopy Coupled Chemometric Algorithms for Rapid Antioxidants Activity Assessment of Chinese Dates (Zizyphus Jujuba Mill.)

Abstract In this work, near-infrared spectroscopy coupled the classical PLS and variable selection algorithms; synergy interval-PLS, backward interval-PLS and genetic algorithm-PLS for rapid measurement of the antioxidant activity of Chinese dates. The chemometric analysis of antioxidant activity assays was performed. The built models were investigated using correlation coefficients of calibration and prediction; root mean square error of prediction, root mean square error of cross-validation and residual predictive deviation (RPD). The correlation coefficient for calibration and prediction sets and RPD values ranged from 0.8503 to 0.9897, 0.8463 to 0.9783 and 1.86 to 4.88, respectively. In addition, variable selection algorithms based on efficient information extracted from acquired spectra were superior to classical PLS. The overall results revealed that near-infrared spectroscopy combined with chemometric algorithms could be used for rapid quantification of antioxidant content in Chinese dates samples.

A Machine Learning Approach for Biomass Characterization

Abstract The aim of this work is to apply and evaluate different chemometric approaches employing several machine learning techniques in order to characterize the moisture content in biomass from data obtained by Near Infrared (NIR) spectroscopy. The approaches include three main parts: a) data pre-processing, b) wavelength selection and c) development of a regression model enabling moisture content measurement. Standard Normal Variate (SNV), Multiplicative Scatter Correction and Savitzky-Golay first (SG1) and second (SG2) derivatives and its combinations were applied for data pre-processing. Genetic algorithm (GA) and iterative PLS (iPLS) were used for wavelength selection. Artificial Neural Network (ANN), Gaussian Process Regression (GPR), Support Vector Regression (SVR) and traditional Partial Least Squares (PLS) regression, were employed as machine learning regression methods. Results shows that SNV combined with SG1 first derivative performs the best in data pre-processing. The GA is the most effective methods for variable selection and GPR achieved a high accuracy in regression modeling while having low demands on computation time. Overall, the machine learning techniques demonstrate a great potential to be used in future NIR spectroscopy applications.

Evaluation of Informative Bands Used in Different PLS Regressions for Estimating Leaf Biochemical Contents from Hyperspectral Reflectance

Partial least squares (PLS) regression models are widely applied in spectroscopy to estimate biochemical components through hyperspectral reflected information. To build PLS regression models based on informative spectral bands, rather than strongly collinear bands contained in the full spectrum, is essential for upholding the performance of models. Yet no consensus has ever been reached on how to select informative bands, even though many techniques have been proposed for estimating plant properties using the vast array of hyperspectral reflectance. In this study, we designed a series of virtual experiments by introducing a dummy variable (Cd) with convertible specific absorption coefficients (SAC) into the well-accepted leaf reflectance PROSPECT-4 model for evaluating popularly adopted informative bands selection techniques, including stepwise-PLS, genetic algorithms PLS (GA-PLS) and PLS with uninformative variable elimination (UVE-PLS). Such virtual experiments have clearly defined responsible wavelength regions related to the dummy input variable, providing objective criteria for model evaluation. Results indicated that although all three techniques examined may estimate leaf biochemical contents efficiently, in most cases the selected bands, unfortunately, did not exactly match known absorption features, casting doubts on their general applicability. The GA-PLS approach was comparatively more efficient at accurately locating the informative bands (with physical and biochemical mechanisms) for estimating leaf biochemical properties and is, therefore, recommended for further applications. Through this study, we have provided objective evaluations of the potential of PLS regressions, which should help to understand the pros and cons of PLS regression models for estimating vegetation biochemical parameters.

NaYF4@Yb,Ho,Au/GO-nanohybrid materials for SERS applications-Pb(II) detection and prediction.

Nowadays, nano hybrid materials (NHMs) with latent applications have been employed in different fields, particularly for sensor applications. Among NHMs, GO based - upconversion NHMs system is an emerging area for the rapid detection of different hazardous materials either as an aptamer based or free-labeled sensing techniques. For the detection of Pb(II) in water, NaYF4@Yb,Ho,Au/GO-NaYF4@Yb,Ho,Au NHMs system was developed. The synthesized NHMs were characterized by XRD, FT-IR, SEM with EDS, TEM and Raman characterization techniques for observation and confirmation. NaYF4@Yb,Ho,Au/GO-NaYF4@Yb,Ho,Au NHMs fabricated sensors were observed to detect and quantify on real-time basis Pb(II) via surface-enhanced-Raman spectroscopy within the dynamic linear range of 98-99%, with detection limits of 1.16 × 10-9 g/mL and 1.15 × 10-8 g/mL obtained respectively for NaYF4@Yb,Ho,Au and GO- NaYF4@Yb,Ho,Au NHMs. The relative standard deviation (RSD) value achieved for both were less than ∼10%, indicative of reproducibility in the quantification results for Pb(II) traces in water when combined with genetic algorithm partial least square (GA-PLS). Results suggest that the GO-NHMs better enhanced the SERS Pb (II) vis-à-vis NaYF4@Yb,Ho,Au. The GO-wrapped NHMs exhibited a further better SERS performance because the heterostructure of GO-NHMs could be potentially useful for SERS-based immunoassay due to the much easier charge transfer between graphene and the metal ions and molecules so the homogeneity of the SERS probe was improved simultaneously (chemical enhancement) by GO-NHMs.

Fast sensing of imidacloprid residue in tea using surface-enhanced Raman scattering by comparative multivariate calibration

This study focused on the fabrication of a rapid, highly sensitive and inexpensive technique for the quantification of imidacloprid residue in green tea, based on surface-enhanced Raman scattering (SERS) using highly roughned surface flower shaped silver nanostructure (as SERS substrate) coupled with the chemometrics algorithm. The basic principle of this method is imidacloprid yielded SERS signal after adsorption on Ag-NF under laser excitation by the electromagnetic enhancement and the intensity of the peak is proportional to the concentration ranging from 1.0 × 103 to 1.0 × 10-4 μg/mL. Among the models used, the GA-PLS (Genetic algorithm-partial least square) exhibited superiority to quantify imidacloprid residue in green tea. The model achieved Rp (correlation coefficient) of 0.9702 with RPD of 4.95% in the test set and RSD for precision recorded up to 4.50%. Therefore, the proposed sensor could be employed to quantify imidacloprid residue in green tea for the safeguarding of quality and human health.

Nondestructive quality assessment of chili peppers using near-infrared hyperspectral imaging combined with multivariate analysis

There is an increasing demand of chili peppers due to their special taste and numerous applications through a number of markets, and high quality is crucial for both producers and customers. This research was aimed to investigate the potential of near-infrared hyperspectral imaging (HSI) for nondestructive quality assessment of chili peppers. Near-infrared HSI in the spectral range of 975–1646 nm was employed to acquire hyperspectral reflectance images of chili peppers. High-performance liquid chromatography and freeze-drying methods were conducted to obtain the reference values of capsaicinoid concentrations and water contents, respectively. Three different variable selection methods with successive projections algorithm (SPA), competitive adaptive reweighted sampling (CARS) and genetic algorithm-partial least squares (GA-PLS) were performed to remove the redundant information and select the optimal wavelengths. Quantitative models including partial least squares (PLS), extreme learning machine (ELM) and least-squares support vector machine (LS-SVM) were then developed to predict the capsaicinoid concentrations and the water content. The results show that the ELM models combined with the SPA method yielded the best prediction performances for the capsaicin and dihydrocapsaicin concentrations, and the water content, with the highest correlation coefficients of prediction (RP) of 0.83, 0.80 and 0.93, respectively. Distribution maps of capsaicin and dihydrocapsaicin concentrations for intact and cut chili peppers were obtained. Finally, classification models for discriminating pungent and non-pungent chili peppers with a classification accuracy of 98.0% were developed. The results demonstrate that near-infrared HSI technique is promising for pepper quality assessment.

Near infrared spectroscopy coupled with chemometric algorithms for predicting chemical components in black goji berries (Lycium ruthenicum Murr.)

Fourier-transform near infrared spectroscopy coupled with chemometric algorithms was applied comparatively for the quantification of chemical compositions in black wolfberry. The compositional parameters, i.e. total flavonoid content, total anthocyanin content, total carotenoid content, total sugar, and total acid were performed for quantification. Model results were evaluated using the correlation coefficients of determination for calibration (R2) and prediction (r2), root-mean-square error of prediction and residual predictive deviation. The findings revealed that the performances of models based on variable selection such as synergy interval-PLS, backward interval-PLS and genetic algorithm-PLS were better than the classical PLS. The performance of the developed models yielded 0.88 ≤ R2 ≤ 0.97, 0.87 ≤ r2 ≤ 0.94 and 1.75 ≤ RPD ≤ 4.00. The overall results showed that the FT-NIR spectroscopy in conjunction with chemometric algorithms could be used for the quantification of the chemical composition of black wolfberry samples.

Synthesized Au NPs@silica composite as surface-enhanced Raman spectroscopy (SERS) substrate for fast sensing trace contaminant in milk.

With increased concerns on milk safety issues, the development of a simple and sensitive method to detect 2,4-dichlorophenoxyacetic acid (2,4-D), a common contaminant in milk, becomes relevant in safeguarding human health threats that results from its consumption. Surface-enhanced Raman spectroscopy (SERS) shows excellent ability for various targets analysis but its usage for rapid and accurate determination of analyte via SERS presents challenges. This study attempted the quantification of 2,4-dichlorophenoxyacetic acid (2,4-D) residue in milk using a novel SERS active substrate- decorated silica films with Au nanoparticles (Au NPs@ silica) coupled to chemometric algorithms. Au NPs@ silica composite was synthesized as a SERS sensor through self-assembly. Thereafter, the SERS spectrum of 2,4-D extract from milk with different concentrations based on the developed SERS sensor was collected and the spectra were analyzed by partial least squares (PLS), and variable selection algorithms - genetic algorithm-PLS (GA-PLS), competitive-adaptive reweighted sampling-PLS (CARS-PLS) and ant colony optimization-PLS (ACO-PLS), to develop quantitative models for 2,4-D prediction. The results obtained showed that the CARS-PLS model gave the optimum result with LOD of 0.01 ng/mL realized and a determination coefficient in the prediction set of (RP) = 0.9836 within a linear range of 10-2 to 106 ng/mL was achieved. Au NPs@ silica SERS sensor combined with CARS-PLS may be employed for rapid quantification of 2,4-D extract from milk towards its quality and safety monitoring.

Docking and QSAR analysis of tetracyclic oxindole derivatives as α-glucosidase inhibitors

The α-glucosidase inhibitors are considered as important agents in drug discovery against diabetes mellitus. Molecular docking and quantitative structure-activity relationship (QSAR) were performed based on a series of tetracyclic oxindole derivatives to elucidate key structural properties affecting inhibitory activity and support the design of new α-glucosidase inhibitors. The molecular docking results demonstrate that at least two hydrogen bonds between Thr681 and Arg676 residues and the oxygen atoms in amid groups have an important role in the optimum binding of inhibitors. In addition, the sum of polar contacts of Arg699, Arg670, Glu792 and Glu301 residues with the α-glucosidase inhibitors have more than one third of total binding free energy. The docked conformations of the inhibitors with the best binding free energy were used to construct QSAR models. As a primary survey and a graphical comparing tool, the partial least squares-discriminant analysis (PLS-DA) technique was successfully employed to classify active and inactive inhibitors. The validated QSAR analysis were performed through genetic algorithm-partial least squares (GA-PLS) and support vector machine (SVM) techniques. The QSAR model reveals that important features of J3D, Mor26 u and HOMA have a high predictive capability (R2p = 0.837, Q2LOO = 0.871, R2LSO = 0.790 and r2m = 0.758) using GA-PLS/SVM strategy. Generally, the suggested QSAR analysis based on classification, docking and GA-PLS/SVM strategy may help suggest chemical scaffold to design novel oxindole derivatives as α-glucosidase inhibitors.

Qualitative and quantitative analysis of chlorpyrifos residues in tea by surface-enhanced Raman spectroscopy (SERS) combined with chemometric models

Surface-enhanced Raman spectroscopy (SERS) combined with chemometric models were employed to develop a rapid, low-cost, and sensitive method for qualitative and quantitative analysis of chlorpyrifos residues in tea. Au@Ag nanoparticles (NPs) with high enhancement factor were synthesized and coupled with chemometric algorithms for SERS measurements. K-nearest neighbors (KNN) classification models gave the best performance model with high classification rates (90.84–100.00%) achieved. For the quantification models for predicting chlorpyrifos contents, the genetic algorithm-partial least squares (GA-PLS) models and synergy interval partial least squares-genetic algorithm (siPLS-GA) models applied to standard normal variate transformation (SNV) preprocessed training and validation data set showed better prediction performances with excellent regression quality (slope = 0.98–1.00), higher correlation coefficient of determination (r2 = 0.96–0.98), and lower root-mean-square error of prediction (RMSEP = 0.29, 0.31) than other quantification models. Paired sample t-test exhibited no statistically significant difference between the reference values determined by GC-MS and the predicted values in most quantification models. The proposed method would be a more effective and powerful tool for classification and determination of chlorpyrifos (CPS) residues in tea samples.

RGO-NS SERS-based coupled chemometric prediction of acetamiprid residue in green tea

Pesticide residue in food is of grave concern in recent years. In this paper, a rapid, sensitive, SERS (Surface-enhanced Raman scattering) active reduced-graphene-oxide-gold-nano-star (rGO-NS) nano-composite nanosensor was developed for the detection of acetamiprid (AC) residue in green tea. Different concentrations of AC combined with rGO-NS nano-composite electro-statically, yielded a strong SERS signal linearly with increasing concentration of AC ranging from 1.0 × 10−4 to 1.0 × 103 μg/mL indicating the potential of rGO-NS nano-composite to detect AC in green tea. Genetic algorithm-partial least squares regression (GA-PLS) algorithm was used to develop a quantitative model for AC residue prediction. The GA-PLS model achieved a correlation coefficient (Rc) of 0.9772 and recovery of the real sample of 97.06%–115.88% and RSD of 5.98% using the developed method. The overall results demonstrated that Raman spectroscopy combined with SERS active rGO-NS nano-composite could be utilized to determine AC residue in green tea to achieve quality and safety.

Near-infrared spectroscopy coupled chemometric algorithms for prediction of antioxidant activity of black goji berries (Lycium ruthenicum Murr.)

In this study, near-infrared (NIR) spectroscopy coupled with partial least-squares (PLS) regression and various efficient variable selection algorithms, synergy interval-PLS (Si-PLS), backward interval PLS (Bi-PLS) and genetic algorithm-PLS (GA-PLS) were applied comparatively for the prediction of antioxidant activity in black wolfberry (BW). The eight assays were used for quantification of antioxidant content. The developed models were assessed using correlation coefficients (R2) of the calibration (Cal.) and prediction (Pre.); root mean square error of prediction, RMSEP; standard Error of Cross-Validation, RMSECV and residual predictive deviation, RPD. The performance of the built model greatly improved by the application of Si-PLS, Bi-PLS and GA-PLS compared with full spectrum PLS. The R2 values determined for calibration and prediction set ranged from 0.8479 to 0.9696 and 0.8401 to 0.9638, respectively. These findings revealed that NIR spectroscopy combined with chemometric algorithms can be used for quantification of antioxidant activity in BW samples.

Catalysis in Chemistry and Biology: Proceedings of the 24th International Solvay Conference on Chemistry

The aim of this work is to apply and evaluate different chemometric approaches employing several machine learning techniques in order to characterize the moisture content in biomass from data obtained by Near Infrared (NIR) spectroscopy. The approaches include three main parts: a) data pre-processing, b) wavelength selection and c) development of a regression model enabling moisture content measurement. Standard Normal Variate (SNV), Multiplicative Scatter Correction and Savitzky-Golay first (SG1) and second (SG2) derivatives and its combinations were applied for data pre-processing. Genetic algorithm (GA) and iterative PLS (iPLS) were used for wavelength selection. Artificial Neural Network (ANN), Gaussian Process Regression (GPR), Support Vector Regression (SVR) and traditional Partial Least Squares (PLS) regression, were employed as machine learning regression methods. Results shows that SNV combined with SG1 first derivative performs the best in data pre-processing. The GA is the most effective methods for variable selection and GPR achieved a high accuracy in regression modeling while having low demands on computation time. Overall, the machine learning techniques demonstrate a great potential to be used in future NIR spectroscopy applications.

Genetic Algorithm Based Wavelengths Selection Coupled with Partial Least Squares for Simultaneous Spectrophotometric Determination of Phosphate and Silicate in Detergent Products

Background: Phosphate and silicate are very important in a variety of compounds such as mineral resources, underground water, detergents and cleaners. Phosphate and silicate in water positively influence the metabolism of plants and animals. They have led to an environmental problem. This problem can have serious consequences because phosphate and silicate are widely utilized in water softening, surfactants and a large number of other applications. Methods: Genetic algorithm (GA) is an appropriate method for selecting wavelengths for PLS (partial least squares) calibration of mixtures which have almost identical spectra without losing prediction capacity using spectrophotometric method. In this paper, we have described a simple, sensitive and selective spectrophotometric method for the simultaneous determination of phosphate and silicate based on the formation of phospho- and silicomolybdenum blue complexes in the presence of ascorbic acid. Results: The complexes of phosphate and silicate with a reagent in the presence of ascorbic acid have been simultaneously determined by genetic algorithm- partial least squares (GA-PLS) although they showed a spectral overlap. The calibration curve was linear over a concentration range of 0.01-2.00 and 0.01-3.80 µgmL-1 for phosphate and silicate, respectively. The root mean square error (RMSE) for phosphate and silicate was 0.372 and 0.301, respectively. Conclusion: The present study indicates that the GA can be a good method for feature selection in spectral data sets. Results show that GA-PLS is an excellent calibration method for the determination of phosphate and silicate. In the end the proposed method was successfully applied for the simultaneous determination of phosphate and silicate in synthetic detergents. Keywords: Phosphate, silicate, spectrophotometry, GA-PLS, detergent products, surfactants.

Target-based drug discovery through inversion of quantitative structure-drug-property relationships and molecular simulation: CA IX-sulphonamide complexes

In this work, a target-based drug screening method is proposed exploiting the synergy effect of ligand-based and structure-based computer-assisted drug design. The new method provides great flexibility in drug design and drug candidates with considerably lower risk in an efficient manner. As a model system, 45 sulphonamides (33 training, 12 testing ligands) in complex with carbonic anhydrase IX were used for development of quantitative structure-activity-lipophilicity (property)-relationships (QSPRs). For each ligand, nearly 5,000 molecular descriptors were calculated, while lipophilicity (logkw) and inhibitory activity (logKi) were used as drug properties. Genetic algorithm-partial least squares (GA-PLS) provided a QSPR model with high prediction capability employing only seven molecular descriptors. As a proof-of-concept, optimal drug structure was obtained by inverting the model with respect to reference drug properties. 3509 ligands were ranked accordingly. Top 10 ligands were further validated through molecular docking. Large-scale MD simulations were performed to test the stability of structures of selected ligands obtained through docking complemented with biophysical experiments.

Advanced QSRR Modeling of Organic Pollutants in Natural Water and Wastewater in Gas Chromatography Time-of-Flight Mass Spectrometry

Water pollution is a major global problem which requires ongoing evaluation and revision of water resource policy at all levels (international down to individual aquifers and wells. It has been suggested that it is the leading worldwide cause of deaths and diseases, and that it accounts for the deaths of more than 14,000 people daily. Genetic algorithm-partial least square (GA-PLS), Kernel partial least square (GA-KPLS) and Levenberg-Marquardt artificial neural network (L-M ANN) techniques were used to investigate the correlation between retention time (RT) and descriptors for 150 organic contaminants in natural water and wastewater which obtained by gas chromatography coupled to high-resolution time-of-flight mass spectrometry (GC-TOF MS). The L-M ANN model gave a significantly better performance than the other models. This indicates that L-M ANN can be used as an alternative modeling tool for quantitative structure–retention relationship (QSRR) studies.

Novel near-infrared spectrum analysis tool: Synergy adaptive moving window model based on immune clone algorithm

This paper presents a novel spectrum analysis tool named synergy adaptive moving window modeling based on immune clone algorithm (SA-MWM-ICA) considering the tedious and inconvenient labor involved in the selection of pre-processing methods and spectral variables by prior experience. In this work, immune clone algorithm is first introduced into the spectrum analysis field as a new optimization strategy, covering the shortage of the relative traditional methods. Based on the working principle of the human immune system, the performance of the quantitative model is regarded as antigen, and a special vector corresponding to the above mentioned antigen is regarded as antibody. The antibody contains a pre-processing method optimization region which is created by 11 decimal digits, and a spectrum variable optimization region which is formed by some moving windows with changeable width and position. A set of original antibodies are created by modeling with this algorithm. After calculating the affinity of these antibodies, those with high affinity will be selected to clone. The regulation for cloning is that the higher the affinity, the more copies will be. In the next step, another import operation named hyper-mutation is applied to the antibodies after cloning. Moreover, the regulation for hyper-mutation is that the lower the affinity, the more possibility will be. Several antibodies with high affinity will be created on the basis of these steps. Groups of simulated dataset, gasoline near-infrared spectra dataset, and soil near-infrared spectra dataset are employed to verify and illustrate the performance of SA-MWM-ICA. Analysis results show that the performance of the quantitative models adopted by SA-MWM-ICA are better especially for structures with relatively complex spectra than traditional models such as partial least squares (PLS), moving window PLS (MWPLS), genetic algorithm PLS (GAPLS), and pretreatment method classification and adjustable parameter changeable size moving window PLS (CA-CSMWPLS). The selected pre-processing methods and spectrum variables are easily explained. The proposed method will converge in few generations and can be used not only for near-infrared spectroscopy analysis but also for other similar spectral analysis, such as infrared spectroscopy.

Approach to chemometrics models by artificial neural network for structure: first applications for estimation retention time of doping agent

A quantitative structure–retention relationship (QSRR), was developed by using the genetic algorithm-partial least square (GA-PLS), Kernel partial least square (GA-KPLS) and Levenberg-Marquardt artificial neural network (L-M ANN) approach for the prediction of the retention time (RT) of the doping agents in urine. The values of the retention time were obtained by using ultra-high-pressure liquid chromatography–quadrupole time-of-flight mass spectrometry (UHPLC-QTOF-MS). A suitable set of the molecular descriptors was calculated and the important descriptors were selected by the aid of the GA-PLS and GA-KPLS. By comparing the results, GA-KPLS descriptors are selected for L-M ANN. Finally a model with a low prediction error and a good correlation coefficient was obtained by L-M ANN. This model was used to predict the RT values of some of doping agents which were not used in the modeling procedure. This is the first research on the QSRR of doping agents against the RT using the GA-PLS, GA-KPLS and L-M ANN model.