Genetic Programming Research Articles

Automatic Design of Energy-Efficient Dispatching Rules for Multi-Objective Dynamic Flexible Job Shop Scheduling Based on Dual Feature Weight Sets

Considering the requirements of the actual production scheduling process, the utilization of the genetic programming hyper-heuristic (GPHH) approach to automatically design dispatching rules (DRs) has recently emerged as a popular optimization approach. However, the decision objects and decision environments for routing and sequencing decisions are different in the dynamic flexible job shop scheduling problem (DFJSSP), leading to different required feature information. Traditional algorithms that allow these two types of scheduling decisions to share one common feature set are not conducive to the further optimization of the evolved DRs, but instead introduce redundant and unnecessary search attempts for algorithm optimization. To address this, some related studies have focused on customizing the feature sets for both routing and sequencing decisions through feature selection when solving single-objective problems. While being effective in reducing the search space, the selected feature sets also diminish the diversity of the obtained DRs, ultimately impacting the optimization performance. Consequently, this paper proposes an improved GPHH with dual feature weight sets for the multi-objective energy-efficient DFJSSP, which includes two novel feature weight measures and one novel hybrid population adjustment strategy. Instead of selecting suitable features, the proposed algorithm assigns appropriate weights to the features based on their multi-objective contribution, which could provide directional guidance to the GPHH while ensuring the search space. Experimental results demonstrate that, compared to existing studies, the proposed algorithm can significantly enhance the optimization performance and interpretability of energy-efficient DRs.

An evolutionary approach to dissolved oxygen mathematical modeling: A case study of the Klamath River

Aquaculture has emerged as a crucial sector in many countries. In the Recirculating Aquaculture System (RAS), Dissolved Oxygen (DO) levels are critical to the health of aquatic animals. As DO sensors are costly, a number of studies have proposed a soft sensor technique utilizing machine learning for estimating DO levels in water. However, the existing research work mainly focuses on black-box approaches, which do not provide numerical analysis between the DO levels and the related parameters. To solve this issue, a sequential Genetic Programming (GP) approach with an evolutionary refinement method is proposed to generate a mathematical expression that represents DO levels in water. In particular, a coarse mathematical model is generated using GP and subsequently fine-tuned using the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). As a study case, the Klamath River dataset is used to generate the model. The evaluation of our proposed method uses datasets from both the Klamath River and the Fanno Creek. Two models are generated in this paper; one model uses six features, while the other only employs three. The results indicate that the model with six features exhibits relatively higher accuracy. However, it is worth noting that a smaller dataset of features is also capable of achieving generalization of the model.

Physics‐constrained symbolic model discovery for polyconvex incompressible hyperelastic materials

AbstractWe present a machine learning framework capable of consistently inferring mathematical expressions of hyperelastic energy functionals for incompressible materials from sparse experimental data and physical laws. To achieve this goal, we propose a polyconvex neural additive model (PNAM) that enables us to express the hyperelastic model in a learnable feature space while enforcing polyconvexity. An upshot of this feature space obtained via the PNAM is that (1) it is spanned by a set of univariate basis functions that can be re‐parametrized with a more complex mathematical form, and (2) the resultant elasticity model is guaranteed to fulfill the polyconvexity, which ensures that the acoustic tensor remains elliptic for any deformation. To further improve the interpretability, we use genetic programming to convert each univariate basis into a compact mathematical expression. The resultant multi‐variable mathematical models obtained from this proposed framework are not only more interpretable but are also proven to fulfill physical laws. By controlling the compactness of the learned symbolic form, the machine learning‐generated mathematical model also requires fewer arithmetic operations than its deep neural network counterparts during deployment. This latter attribute is crucial for scaling large‐scale simulations where the constitutive responses of every integration point must be updated within each incremental time step. We compare our proposed model discovery framework against other state‐of‐the‐art alternatives to assess the robustness and efficiency of the training algorithms and examine the trade‐off between interpretability, accuracy, and precision of the learned symbolic hyperelastic models obtained from different approaches. Our numerical results suggest that our approach extrapolates well outside the training data regime due to the precise incorporation of physics‐based knowledge.

Prediction of seismic-induced bending moment and lateral displacement in closed and open-ended pipe piles: A genetic programming approach

Ensuring the reliability of pipe pile designs under earthquake loading necessitates an accurate determination of lateral displacement and bending moment, typically achieved through complex numerical modeling to address the intricacies of soil-pile interaction. Despite recent advancements in machine learning techniques, there is a persistent need to establish data-driven models that can predict these parameters without using numerical simulations due to the difficulties in conducting correct numerical simulations and the need for constitutive modelling parameters that are not readily available. This research presents novel lateral displacement and bending moment predictive models for closed and open-ended pipe piles, employing a Genetic Programming (GP) approach. Utilizing a soil dataset extracted from existing literature, comprising 392 data points for both pile types embedded in cohesionless soil and subjected to earthquake loading, the study intentionally limited input parameters to three features to enhance model simplicity: Standard Penetration Test (SPT) corrected blow count (N60), Peak Ground Acceleration (PGA), and pile slenderness ratio (L/D). Model performance was assessed via coefficient of determination (R2), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE), with R2 values ranging from 0.95 to 0.99 for the training set, and from 0.92 to 0.98 for the testing set, which indicate of high accuracy of prediction. Finally, the study concludes with a sensitivity analysis, evaluating the influence of each input parameter across different pile types.

Cooling load prediction of a double-story terrace house using ensemble learning techniques and genetic programming with SHAP approach

Since the cooling systems used in buildings in hot climates account for a significant portion of the energy consumption, it is very important for both economy and environment to accurately predict the cooling load and consider it in building designs. This study aimed to maximize energy efficiency by appropriately selecting the features of a building that affect its cooling load. To this end, data-driven, accurate, and accessible tools were developed that enable the prediction of the cooling load of a building by practitioners. The study involves simulating the energy consumption of a mid-rise, double-story terrace house in Malaysia using building information modeling (BIM) and estimating the cooling load using ensemble machine learning models and genetic programming. Categorical Boosting (CatBoost), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Random Forest (RF) models have been developed and made available as an online interactive graphical user interface on the Streamlit platform. Furthermore, the symbolic regression technique has been utilized to obtain a closed-form equation that predicts the cooling load. The dataset used for training the predictive models comprised 94,310 data points with 10 input variables and the cooling load as the output variable. Performance metrics such as the coefficient of determination (R2), root mean squared error (RMSE), and mean absolute error (MAE) were used to measure the predictive model performances. The results of the machine learning models indicated successful prediction, with the CatBoost model achieving the highest score (R2 = 0.9990) among the four ensemble models and the predictive equation. The SHAP analysis determined the aspect ratio of the building as the most impactful feature of the building.

Investigating the etiology of non-surgical hypoparathyroidism: insights from a sponsored genetic testing program

Investigating the etiology of non-surgical hypoparathyroidism: insights from a sponsored genetic testing program

Generation of 'semi-guided' cortical organoids with complex neural oscillations.

Temporal development of neural electrophysiology follows genetic programming, similar to cellular maturation and organization during development. The emergent properties of this electrophysiological development, namely neural oscillations, can be used to characterize brain development. Recently, we utilized the innate programming encoded in the human genome to generate functionally mature cortical organoids. In brief, stem cells are suspended in culture via continuous shaking and naturally aggregate into embryoid bodies before being exposed to media formulations for neural induction, differentiation and maturation. The specific culture format, media composition and duration of exposure to these media distinguish organoid protocols and determine whether a protocol is guided or unguided toward specific neural fate. The 'semi-guided' protocol presented here has shorter induction and differentiation steps with less-specific patterning molecules than most guided protocols but maintains the use of neurotrophic factors such as brain-derived growth factor and neurotrophin-3, unlike unguided approaches. This approach yields the cell type diversity of unguided approaches while maintaining reproducibility for disease modeling. Importantly, we characterized the electrophysiology of these organoids and found that they recapitulate the maturation of neural oscillations observed in the developing human brain, a feature not shown with other approaches. This protocol represents the potential first steps toward bridging molecular and cellular biology to human cognition, and it has already been used to discover underlying features of human brain development, evolution and neurological conditions. Experienced cell culture technicians can expect the protocol to take 1 month, with extended maturation, electrophysiology recording, and adeno-associated virus transduction procedure options.

A New Road Extraction Method from Satellite Images Using Genetic Programming

A New Road Extraction Method from Satellite Images Using Genetic Programming

Expanded targeted preconception screening panel in Israel: findings and insights

BackgroundWe aimed to analyse the efficacy and added value of a targeted Israeli expanded carrier screening panel (IL-ECSP), beyond the first-tier test covered by the Israeli Ministry of Health (IMOH)...

Extremely low frequency–electromagnetic fields promote chondrogenic differentiation of adipose-derived mesenchymal stem cells through a conventional genetic program

Progressive cartilage deterioration leads to chronic inflammation and loss of joint function, causing osteoarthritis (OA) and joint disease. Although symptoms vary among individuals, the disease can cause severe pain and permanent disability, and effective therapies are urgently needed. Human Adipose-Derived Stem Cells (ADSCs) may differentiate into chondrocytes and are promising for treating OA. Moreover, recent studies indicate that electromagnetic fields (EMFs) could positively affect the chondrogenic differentiation potential of ADSCs. In this work, we investigated the impact of EMFs with frequencies of 35 Hertz and 58 Hertz, referred to as extremely low frequency-EMFs (ELF–EMFs), on the chondrogenesis of ADSCs, cultured in both monolayer and 3D cell micromasses. ADSC cultures were daily stimulated for 36 min with ELF–EMFs or left unstimulated, and the progression of the differentiation process was evaluated by morphological analysis, extracellular matrix deposition, and gene expression profiling of chondrogenic markers. In both culturing conditions, stimulation with ELF–EMFs did not compromise cell viability but accelerated chondrogenesis by enhancing the secretion and deposition of extracellular matrix components at earlier time points in comparison to unstimulated cells. This study showed that, in an appropriate chondrogenic microenvironment, ELF–EMFs enhance chondrogenic differentiation and may be an important tool for supporting and accelerating the treatment of OA through autologous adipose stem cell therapy.

Genetic Programming and Reinforcement Learning on Learning Heuristics for Dynamic Scheduling: A Preliminary Comparison

Genetic Programming and Reinforcement Learning on Learning Heuristics for Dynamic Scheduling: A Preliminary Comparison

A needle in a haystack? The impact of a targeted epilepsy gene panel in the identification of a treatable but rapidly progressive metabolic epilepsy: CLN2 disease.

Neuronal ceroid lipofuscinoses (NCL) are a group of autosomal recessive, inherited, lysosomal, and neurodegenerative diseases that causes progressive dementia, seizures, movement disorders, language delay/regression, progressive visual failure, and early death. Neuronal ceroid lipofuscinosis type 2 (CLN2), caused by biallelic pathogenic variants of the TPP1 gene, is the only NCL with an approved targeted therapy. The laboratory diagnosis of CLN2 is established through highly specific tests, leading to diagnostic delays and eventually hampering the provision of specific treatment for patients with CLN2. Epilepsy is a common and clinically-identifiable feature among NCLs, and seizure onset is the main driver for families to seek medical care. To evaluate the results of the Latin America Epilepsy and Genetics Program, an epilepsy gene panel, as a comprehensive tool for the investigation of CLN2 among other genetic causes of epilepsy. A total of 1,284 patients with epilepsy without a specific cause who had at least 1 symptom associated with CLN2 were screened for variants in 160 genes associated with epilepsy or metabolic disorders presenting with epilepsy through an epilepsy gene panel. Variants of the TPP1 gene were identified in 25 individuals (1.9%), 21 of them with 2 variants. The 2 most frequently reported variants were p.Arg208* and p.Asp276Val, and 2 novel variants were detected in the present study: p.Leu308Pro and c.89 + 3G > C Intron 2. The results suggest that these genetic panels can be very useful tools to confirm or exclude CLN2 diagnosis and, if confirmed, provide disease-specific treatment for the patients.

Fuzzy inference systems for discretionary lane changing decisions: Model improvements and research challenges

Lane Changing Decision Model (LCDM) is a critical component in semi- and fully-automated driving systems. Recent research has found that Fuzzy Inference System (FIS) is a promising approach to implementing LCDMs. To improve the FIS’s performance, this research reviewed the challenges in the development of a FIS model to make the yes,no decisions in discretionary lane changes. The FIS model was revised to bring its fuzzy inference rules more consistent with the fuzzy membership functions, and its composition and defuzzification methods more in line with the classical fuzzy logic theory. An equitable test data set with approximately equal number of yes,no data points was assembled from the same NGSIM data used in the past research. The test results proved that: (1) A LCDM’s performance was dependent on how the yes,no decisions in the test data set were manually labeled; and (2) Separating the fuzzy inference rules into a yes group and a no group and compute the results separately yielded potentially better decision accuracy. Furthermore, The Genetic Expression Programming Model (GEPM) performed better than the improved FIS-based model. The findings led the authors to suggest two possible research directions: (1) Add the subject vehicle’s speed as an input to the LCDM and redesign the decision-making model; and (2) Construct models for congested and uncongested traffic separately. The authors further suggested the use of instrumented vehicles to collect a set of high-fidelity lane changing data in the naturalistic driving environment. (word count: 246)

Unacylated Ghrelin Protects Against Muscle Wasting, Mitochondrial Dysfunction, and Neuromuscular Junction Disruption in Tumor Bearing Mice

Background: Cancer cachexia is a complex metabolic syndrome that severely impacts mobility, treatment strategies, and life quality of the patients. However, no treatments are available to mitigate the debilitating consequences of cancer cachexia. Ghrelin is a hormone released from the stomach that increases appetite upon acylation by binding to GHSR1a receptors in the brain. Because of its orexigenic effects and increases in body weight, ghrelin GHSR1a receptor analogs were tested as clinical trials to mitigate muscle wasting and loss of strength in cancer patients in several countries. The analogs (i.e., Anamorelin) increased body weight of the non-small lung cancer patients in European countries and Japan, but failed to increase strength, which is the critical component in cancer cachexia. Mainly due to the lack of improvement in strength and patients’ quality of life, Anamorelin was not approved for clinical use in Europe. While the lack of increases in strength is, at least in part, associated with the adipogenic effects of acylated ghrelin, unacylated ghrelin directly improves satellite cell renewal, muscle growth, and mitochondrial function independent of GHSR1a receptor activation and increases in appetite. Unacylated ghrelin is the predominant population in circulation comprising proximately 90-95% of total ghrelin in humans and rodents. Hypothesis: Unacylated ghrelin projects against muscle wasting, loss of strength, and metabolic dysfunction of tumor bearing mice. Methods and Results: Supplementing the cancer cell conditioned medium (CM) with unacylated ghrelin preserved myotube size. Unacylated ghrelin also diminished the cancer cell CM-induced upregulation of MuRF-1, muscle specific ubiquitin ligase. Interestingly, unacylated ghrelin normalized transcriptional upregulation of denervation markers (i.e., sarcolipin, Runx1, GADD45a) in myotubes treated with CM. To further investigate the therapeutic effects of unacylated ghrelin in vivo, we inoculated lung cancer cell line (LL2, ATCC®) in the flank of 4 months old male mice. Two weeks after the injection prior to development of cachexia, mice were provided with unacylated ghrelin containing drinking water or untreated water as control. Tumor bearing (TB) mice treated with unacylated ghrelin showed 30-40% increases in gastrocnemius and quadriceps weights, while soleus muscle mass was unchanged. In addition to the muscle quantity, unacylated ghrelin increased force generating capacity was also improved by unacylated ghrelin. Notably, unacylated ghrelin increased mitochondrial oxygen consumption rate and protected against neuromuscular junction disruption and denervation in TB control mice. Conclusion: Unacylated ghrelin protects against cancer cachexia by targeting multiple risk factors in skeletal muscle wasting, including protein balance, metabolic alterations, and neuromuscular junction disruption. Our findings show protective effects of unacylated ghrelin against cachexia and neuromuscular impairment in tumor bearing mice, indicating its therapeutic potential in cancer patients. NIA R00 064143 to Ahn B. Pilot funds from the Center for Redox Biology in Medicine, and Cancer Genetics and Metabolism program at the Wake Forest Baptist Comprehensive Cancer Center. This is the full abstract presented at the American Physiology Summit 2024 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.

Oscillatory control of embryonic development.

Proper embryonic development depends on the timely progression of a genetic program. One of the key mechanisms for achieving precise control of developmental timing is to use gene expression oscillations. In this Review, we examine how gene expression oscillations encode temporal information during vertebrate embryonic development by discussing the gene expression oscillations occurring during somitogenesis, neurogenesis, myogenesis and pancreas development. These oscillations play important but varied physiological functions in different contexts. Oscillations control the period of somite formation during somitogenesis, whereas they regulate the proliferation-to-differentiation switch of stem cells and progenitor cells during neurogenesis, myogenesis and pancreas development. We describe the similarities and differences of the expression pattern in space (i.e. whether oscillations are synchronous or asynchronous across neighboring cells) and in time (i.e. different time scales) of mammalian Hes/zebrafish Her genes and their targets in different tissues. We further summarize experimental evidence for the functional role of their oscillations. Finally, we discuss the outstanding questions for future research.

Using AI to identify left ventricular ejection fraction from the ECG: The SOLOMAX (SOciaL NetwOrk of MedicAl Experiences) project

Aim: The interest in machine learning-based algorithms in the cardiovascular field is rapidly growing, especially for diagnostic and prognostic purposes. Recent evidence has demonstrated that certain electrocardiographic (ECG) parameters are predominantly associated with systolic function, estimated as left ventricular ejection fraction (LVEF) by echocardiography, albeit with still relatively low accuracy. Consequently, this study aims to develop an AI-based model capable of predicting LVEF from ECG data in an Italian population. Methods: Within the SOLOMAX project, we collected paired ECG-Echocardiography exams from 105 patients (64.82±16.02y;62.86%male). Precisely, we excluded patients with atrial fibrillation at the time of the ECG, PMK or electrostimulated rhythm, valve prostheses, previous cardiac surgery, O2 therapy or COPD, previous ablation or invasive electrophysiology procedures, currently hospitalized for Takotsubo or ACS, heart failure exacerbation, inotropic therapy, ACS over the last 3 months. We recorded anthropometric, clinical, biochemical, ECG, and Echocardiography parameters. The collected data was studied using AI-based techniques to create a new model to predict LVEF from ECG. Using an approach based on evolutionary algorithms, genetic programming was used. This approach solves a symbolic regression problem through genetic algorithms and provides a mathematical model of the relationship between ECG parameters and LVEF. The formula obtained was then used to build a simple explainable classifier, which provides a global interpretation of the link between ECG parameters and LVEF. Results: The performance of the proposed approach and the reliability of the results were assessed using the k-fold cross-validation method and by estimating standard metrics derived from the confusion matrix associated with a binary classifier, that is, accuracy, sensitivity, specificity, precision, and F-Measure. The proposed approach consistently demonstrated its ability to distinguish patients with preserved LVEF from those with reduced LVEF. Each metric averaged across all experiments scored approximately 95%. Furthermore, in the expression generated by the AI model, the axes of the P, QRS, and T waves play a prominent role, as they are likely to provide a better interpretation of the three-dimensional cardiac geometry and, consequently, cardiac function. Conclusions: AI applied to ECG data can be used to create cost-effective diagnostic and predictive tools for assessing LVEF. Indeed, the obtained formula highlights the relationship between ECG parameters and LVEF, as well as its complexity, which can aid in detecting heart diseases.

Modeling and future projection of streamflow and sediment yield in a sub-basin of Euphrates River under the effect of climate change

ABSTRACT Recognizing the differential impacts of climate change across geographical scales, this study emphasizes the importance of statistical downscaling. Using Gene Expression Programming (GEP) and Linear Genetic Programming (LGP), statistical downscaling transforms broad climate trends into region-specific insights. This allowed for detailed analyses of anticipated changes in sediment yield and discharge within a Euphrates River sub-basin in Türkiye using large-scale variables from the CanESM2 model. The dataset is divided into calibration (1970–1995) and validation (1996–2005) periods. To assess the models’ accuracy, statistical measures such as RMSE, MAE, NSE, and R were used. The analysis revealed that LGP outperformed GEP in both discharge and sediment yield during validation, with RMSE = 51.79 m3/s and 4,325.66 tons/day, MAE = 27.14 m3/s and 1,593.34 tons/day, NSE = 0.684 and 0.627, and R = 0.841 and 0.788, respectively. However, when simulating future periods based on the observed period (2006–2020), the GEP model was superior to LGP under RCP2.6, RCP4.5, and RCP 8.5 scenarios from CanESM2. In 2021–2100, models suggest a moderate decrease in discharge and sediment yield, indicating potential shifts in the basin's hydrodynamics. These changes could disrupt hydropower generation, challenge water management practices, and alter riverine ecosystems. The results necessitate a thorough assessment of potential ecological consequences.

Evolution in Patterns of Urban Water Consumption Accompanying Socio-Economic Development

To explore the evolution in patterns of urban water consumption (UWC) accompanying socio-economic development, historical data on socio-economic conditions and water consumption in developed cities worldwide were collected. stages of evolution and patterns in UWC were identified through Pettitt tests. Through correlation analysis, the main socio-economic indicators influencing UWC were identified, and their threshold values for different trends in the evolution of UWC were statistically analyzed. The use of these patterns of evolution for UWC prediction was explored taking Changsha, China as a case study. Results indicate the following: (1) UWC typically increases in the early stage of development and then stabilizes or decreases later; (2) when UWC stabilizes or decreases, socio-economic indicators, namely per capita GDP, tertiary industry’s contribution to GDP, and urbanization rate, range from [USD 10,000, 60,000], [60%, 80%], and [85%, 95%], respectively, entering or approaching the developed economy stage. (3) Using patterns in this evolution, Changsha’s UWC is predicted to increase until the turning point, which will occur between 2027 and 2035 (earlier than that predicted using genetic programming) and then decrease earlier than that of Hunan Province or the entirety of China. Predicted values are lower than those obtained using the quota method. This study provides a new perspective on relationships between UWC and socio-economic conditions, supporting water resources planning.

Adaptive energy management strategy for solar energy harvesting IoT nodes by evolutionary fuzzy rules

This study explores the integration of genetic programming (GP) and fuzzy logic to enhance control strategies for Internet of Things (IoT) nodes across varied locations. It is introduced a novel methodology for designing a fuzzy-based energy management controller that autonomously determines the most suitable controller structure and inputs. This approach is evaluated using a solar harvesting IoT model that leverages historical solar irradiance data, highlighting the methodology’s potential for diverse geographical applications and compatibility with low-performance microcontrollers. The findings demonstrate that the integration of GP with designed fitness function enables the dynamic learning and adaptation of control strategies, optimizing system behavior based on historical data. The experimental model showcases an ability to efficiently use historical datasets to derive optimal control strategies, with the fitness metric indicating consistent improvement throughout the learning phase. The results indicate that useful control strategies learned at a certain location may outperform a locally-trained control strategies and can be successfully re-applied in other locations.

A bZIP type transcription factor from Picrorhiza kurrooa modulates the expression of terpenoid biosynthesis genes in Nicotiana benthamiana

bZIP family of transcription factors are DNA-binding proteins which bind ACGT motifs in the upstream promoter region of genes. bZIP proteins play significant roles in regulation of plant growth, secondary metabolism, development, defense and response to various abiotic and biotic stresses. Up to now, there is no report on molecular and functional characterization of bZIP transcription factors in Picrorhiza kurrooa, a medicinally important endangered herb of north western Himalayan region. The species is widely used in several commercially available drug formulations like livomap, livocare and livplus for treatment of liver disorders. The medicinal properties of P. kurrooa are due to bioactive picrosides, synthesized by isoprenoid pathway. The promoters of regulatory genes viz., Pkdxs and Pkhmgr of picroside biosynthesis pathway have been reported to contain ACGT cis-acting elements (bzip motifs) at different locations. In this study, the functionality of ACGT motifs was confirmed by electrophoretic mobility shift assay (EMSA). Full length PkbZIP gene was cloned through RACE method. In silico analysis revealed PkbZIP contained a conserved signature bZIP domain (N-x7-R/K-x9-L-x6-L-x6-L). Transcriptional activation and subcellular localization assay confirmed nuclear localization of PkbZIP. EMSA with purified PkbZIP confirmed in vitro binding of PkbZIP to ACGT elements in Pkdxs and Pkhmgr promoters. The expression pattern of PkbZIP positively correlated with the picroside (active medicinal component of the plant) content in different tissues, at 15 °C and under light conditions. Transient overexpression in Nicotiana benthamiana suggested a role of PkbZIP in terpenoid biosynthesis. PkbZIP might prove as potential candidate gene to be used in genetic engineering program for enhancing the picroside content in P. kurrooa.