Qualification: 
Ph.D, MPhil, MSc
os_deepa@cb.amrita.edu

Dr. Deepa Gopakumar O. S. currently serves as Associate Professor in the Department of Mathematics, School of Engineering, Coimbatore. Her areas of research include Statistical Quality Control, Statistical Methods in Bio Informatics.

Publications

Publication Type: Journal Article

Year of Publication Title

2018

Ani R., Manohar, R., Anil, G., and Dr. Deepa Gopakumar O. S., “Virtual Screening of Drug Likeness using Tree Based Ensemble Classifier”, Biomedical and Pharmacology Journal, vol. 11, pp. 1513-1519, 2018.[Abstract]


In earlier years, the Drug discovery process took years to identify and process a Drug. It takes a normal of 12 years for a Drug to travel from the research lab to the patient. With the introduction of Machine Learning in Drug discovery, the whole process turned out to be simple. The utilization of computational tools in the early stages of Drug development has expanded in recent decades. A computational procedure carried out in Drug discovery process is Virtual Screening (VS). VS are used to identify the compounds which can bind to a Drug target. The preliminary process before analyzing the bonding of ligand and drug protein target is the prediction of drug likeness of compounds. The main objective of this study is to predict Drug likeness properties of Drug compounds based on molecular descriptor information using Tree based ensembles. In this study, many classification algorithms are analyzed and the accuracy for the prediction of drug likeness is calculated. The study shows that accuracy of rotation forest outperforms the accuracy of other classification algorithms in the prediction of drug likeness of chemical compounds. The measured accuracies of the Rotation Forest, Random Forest, Support Vector Machines, KNN, Decision Tree and Naïve Bayes are 98%, 97%, 94.8%, 92.8%, 91.4%, 89.5% respectively

More »»

2018

G. Shruthi and Dr. Deepa Gopakumar O. S., “Average run length for exponentiated distribution under truncated life test”, International Journal of Mechanical Engineering and Technology, vol. 9, pp. 1180-1188, 2018.[Abstract]


Statistical quality control discusses about the screening and preserving the quality of products and services. Control charts are used widely in manufacturing industry. The acceptance sampling plans for truncated life tests are regularly practiced to regulate the sample size from a lot under inspection. Traditional control chart needs more samples for testing the quality of the product. In this paper, failure time of a product follows non-normal distributions like Exponentiated Gamma Distribution, Exponentiated Lomax Distribution and Beta Weibull Distribution. The defectives that are found are put into truncated life test and upper and lower control limits are computed. Control charts using these distributions are designed to monitor the mean shift by detecting the number of unsuccessful products at a definite time. The probability for the in-control process and out of control process are estimated from the sample data. The Average Run Length (ARL) allows an assessment to be considered for various screening policies and based on the various mean shift, the ARL values are computed for various distributions and compared. © IAEME Publication.

More »»

2016

Dr. Deepa Gopakumar O. S. and Ani R., “Docking studies of Hemochromatosis protein with various compounds of the medicinal plants”, Journal of Chemical and Pharmaceutical Sciences, no. 4, pp. 10-13, 2016.[Abstract]


Hemochromatosis is a genetic disorder which leads to the accretion of iron in parenchymal organs leading to organ toxicity. Normal absorption of iron from daily food is 10% whereas people with hemochromatosis diseases can absorb iron four times more than the normal absorption. There is no proper medication and clinically proved medicines have side effects. Hence an alternative methods is the extraction of bioactive compounds from the medicinal plants to recognize the novel target. 17 compounds from 13 medicinal plants with 81 properties were collected from database. 81 properties with certain specific conditions were checked for accuracy using various machine learning techniques. The compounds that satisfies the pharmacological properties are docked with the mutated protein of hemochromatosis. The free energy and the interaction analysis were also discussed.

More »»

2016

Ani R. and Dr. Deepa Gopakumar O. S., “Rotation Forest Ensemble Algorithm for the Classification of Phytochemicals from the Medicinal Plants”, Journal of Chemical and Pharmaceutical Sciences, no. 4, pp. 14-17, 2016.[Abstract]


Drug Discovery from medicinal plants is an important area in current research and has been providing important source of new drug leads. Plant extracts are proved as main source for many drugs. A major part of traditional therapy uses plant extracts or the associated active principles. Many of the traditional medicines are made as a result of applying some small synthetic modifications of naturally obtained substances. But most of the modern medicines are using synthetic substances instead of natural substances obtained from medicinal plants. Few machine
learning predictive algorithms are applied to classify the compounds to the defined classes and the accuracies of different classification algorithms are analyzed. The present study shows the significance of Rotation Forest ensemble algorithms in the classification of medicinal plant compounds. The other algorithms analyzed are Decision Tree, Random Forest and Naive Bayes. The Random Forest tree based ensemble outperformed the other algorithms in this study.

More »»

2016

Dr. Deepa Gopakumar O. S. and Ani R., “Expectation - Maximization algorithm for protein - Ligand complex of HFE gene”, Journal of Chemical and Pharmaceutical Sciences, vol. 2016, pp. 6-9, 2016.[Abstract]


In the field of pharmaceutical sciences and biomedicine, the issue of protein stabilization presumes meticulous importance. It plays a significant role in purification, formulation, and storage. Suitably folded proteins are usually stable during expression and purification. The interaction between ligands and proteins generally produces changes in protein thermal stability with changes in the midpoint denaturation temperature, enthalpy of unfolding, and heat capacity. The stability of eleven mutations of the proteins corresponding to HFE gene are identified using Random forest and Support vector machine. Various parameters like Half-life period, aliphatic index and GRAVY are computed using online webservers. Based on the machine learning techniques and the computed parameters, the ligands for HFE proteins are obtained. The contact surface area between ligand atom and protein atoms are also identified. The expectation - maximization algorithm was done on the contact surface area to test whether there exist any change in the destabilizing contact between the ligand atom and protein atoms.

More »»

2015

Dr. Deepa Gopakumar O. S., Nallamalli, S., L. Naik, N. Singh, and Teja, G. Venkata Sa, “Mathematical Model for Transmission of Ebola Virus”, Procedia Computer Science, vol. 48, 2015.

2015

Dr. Deepa Gopakumar O. S., “Optimal Production Policy for the Design of Green Supply Chain Model ”, International Journal of Applied Engineering Research, vol. 10, 2015.

2015

Dr. Deepa Gopakumar O. S., “Application of Acceptance Sampling Plan in Green Design and Manufacturing ”, International Journal of Applied Engineering Research , vol. 10, 2015.

2013

Dr. Deepa Gopakumar O. S., “Analysis of chain sampling plan for exponential family ”, Mathematical Sciences International Research Journal, 2013.

2011

K. U. Radhagayathri, Mohandas, V. P., Subeesh, T., Dr. Deepa Gopakumar O. S., P. K. Krishnan Namboori, and Ramachandran, K. I., “Computational modeling and simulation of nanomolecular switch for Alzheimer's disease (A gene silencing technique)”, International Journal of Nanoscience, vol. 10, pp. 319-322, 2011.[Abstract]


The design of artificial gene regulatory networks has paved way for the construction of therapeutic gene circuits that would find application in next generation gene therapy approaches. The main challenge in such designs is in selecting the appropriate genetic components to make up the circuit in order to produce the anticipated or desired behavior. To eliminate this complexity, computational simulation tools are used to guide circuit design. This involves selection and genetic modification of components, until the required system behavior is achieved. In this work, we have designed a model for a synthetic nanogene network. The gene expression involved in diseases caused by mutation, like cancer, Alzheimer's disease (AD), etc., can be effectively controlled by "Nanogene silencing genetic switch". Here, we have considered the case of Alzheimer's disease. Genes that are responsible for genetic AD are APP, PSEN1, and PSEN2. Antimutagenic study of phytochemicals has been carried out with the common AD causing APP gene mutation. By our computational analysis, phytochemicals like curcumin, eugenol, and limonene have been identified as "molecular switch" devices to prevent mutation of AD gene. All these chemicals are found to be strongly interacted with the AD gene without making changes to the remaining part of the sequence. Among the analyzed chemicals, curcumin is found to be most interacting with AD gene protecting it from mutation. Hence, inclusion of curcumin-containing food items in our menu would prevent Alzheimer's disease to a large extent. This technique is a type of gene therapy leading to silencing of the mutated part of the gene and preventing mutation. © 2011 World Scientific Publishing Company.

More »»

Publication Type: Conference Paper

Year of Publication Title

2016

Dr. Deepa Gopakumar O. S., Ani R., Sasi, G., and Sankar, R., “Decision Support system for diagnosis and prediction of Chronic Renal Failure using Random Subspace Classification”, in 2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), Jaipur, India, 2016.[Abstract]


Chronic Renal Failure (CRF) is one of the major disease which affect the human life. The stages of CRF start with loss of renal functions and gradually it leads to complete failure of all kidney functions. This disease is fatal at its end stage unless a replacement of kidney or a dialysis process which is an artificial filtering mechanism is not done. So an early prediction of disease is very important to save the human life. Machine learning is a part of artificial intelligence that uses a variety of techniques to learn from complex dataset. Machine learning techniques are widely used in medical field for disease prediction and prognosis. The objective of this work is to develop a clinical decision support system using machine learning techniques. In this paper first the classification techniques like neural network based back propagation (BPN), probability based Naive Bayes, LDA classifier, lazy learner K Nearest Neighbor (KNN), tree based decision tree, and Random subspace classification algorithms are analyzed. The accuracy of each algorithm found is 81.5%, 78%, 76%, 90%, 93% and 94% respectively on a dataset collected from UCI repository which contains 25 attributes and 400 instances. From the results obtained, the algorithm which gave better result was used for the developing the Clinical Decision Support System.

More »»

2016

Ani R., Augustine, A., Akhil, N. C., and Dr. Deepa Gopakumar O. S., “Random Forest Ensemble Classifier to Predict the Coronary Heart Disease Using Risk Factors”, in Proceedings of the International Conference on Soft Computing Systems, 2016.[Abstract]


Heart diseases are the major cause of death in today’s modern age. Coronary heart disease is one among them. This disease attacks the normal person instantly. Proper diagnosis and timely attention to the patients reduce mortality rate. Proper diagnosis has become a challenging task for the medical practitioners. The cost involved in the immediate treatment or intervention methods are also very expensive. Early diagnosis of the disease using mining of medical data prevents the inattention of occurrence of sudden CHD events. Today, almost all hospitals are using hospital information system and it has huge volume of patient records. This study results in the development of a decision support system using machine intelligence techniques applied on the medical records stored in hospital databases. Classification algorithms are used to evaluate the accuracy of the early prediction of coronary heart events. The classification techniques analyzed are K-nearest neighbor, decision tree-C4.5, Naive Bayes, and the random forest. The accuracy of each technique is found to be 77, 81, 84, and 89 %, respectively. In this study 10-fold cross-validation method is used to measure the unbiased estimate of these prediction models.

More »»

2014

Dr. Deepa Gopakumar O. S., “Selection of Sampling plans with double binomial as prior distribution”, in International conference on Mathematical Science , 2014.

2009

V. K Gopal, Premkumar, P., Krishnan, N. P. K., Sabarish Narayanan B., and Dr. Deepa Gopakumar O. S., “In-Silico Modeling and Simulation of Magnetic Nanoparticles for the Biological Cell Isolation Technique”, in International Conference on Artificial Neural Networks (ICANN), IIT, Guwahati, 2009.