http://dkr.cdri.res.in:8080/xmlui/handle/123456789/28 2026-04-19T13:41:22Z 2026-04-19T13:41:22Z Abbas, M Mandal, S K Srivastava, M http://dkr.cdri.res.in:8080/xmlui/handle/123456789/512 2010-02-10T20:34:05Z 2006-01-01T00:00:00Z

Statistical and Mathematical Methods in Bioinformatics- An Overview Abbas, M; Mandal, S K; Srivastava, M Bioinformatics is the buzzword these days. Bioinformatics may be considered as the study of information flow within biology and medicine. The first flow is the flow of information from DNA code to biological function. Second flow is the flow of 'in formation in the design and analysis of experiments. Studies in the first flow include methods for sequence alignment, gene finding, RNA expression, protein expression, prediction of protein 3D structure, population genetics, and modelling of genetic network. The second flow begins with a hypothesis (drawn by scanning molecular biology databases), followed by a plan to collect data, execution of an experiment and analysis of the results. Bioinformatics has gained prominence recently because biologists can now collect huge amount of data by using high through put techniques. Large number of relevant databases are available on the internet and a number of software tools and techniques are available for data analysis. Algorithms which have been implemented in these software are based on a number of mathematical and statistical frameworks. Prominent among these are : Dynamic programming, Discriminant analysis, Neural network, Markov model, Multivariate analysis, and it couple of Machine learning techniques.

2006-01-01T00:00:00Z Srivastava, M Abbas, M Mandal, S K http://dkr.cdri.res.in:8080/xmlui/handle/123456789/511 2010-02-10T20:33:25Z 2006-01-01T00:00:00Z

Statistical evaluation of biological activity of plant based products: Anti-diabetic screening Srivastava, M; Abbas, M; Mandal, S K In classical approach of drug discovery involving medicinal plants, large number of extracts are prepared by using different parts of plants and solvents. The biological activity of these extracts are assessed using some standard test system. Extracts which are found active are fractionated to get different fractions. These fractions are assessed using the suitable test system. Fractions which are found active are processed further to get pure compounds, which are again tested for bioactivity. At every stage (extracts, fractions and pure compounds), lot of test data is generated. Statisticians play an Important role in evaluating such 4ata to get promising compounds. If the activity of any compound is over- estimated, then chances of it getting dropped at preclinical or clinical level will be higher, and if under-estimated and not pursued further, we may loose a good drug. Depending upon the disease and observations of response, a suitable method is applied to ascertain the significantly active compounds. For example, to get a lead against diabetes one has to select a suitable animal model such as rat, mice, etc. N number of animals are considered in two groups. Both group animals are loaded with glucose. One randomly taken group is administered the compound under test, while the other one remains untreated and acts as control. The serum glucose level of the animals is measured at different predecided time points.

2006-01-01T00:00:00Z Abbas, M Mandal, S K Srivastava, M http://dkr.cdri.res.in:8080/xmlui/handle/123456789/510 2010-02-10T20:32:17Z 2006-01-01T00:00:00Z

DNA Microarray Data Management and Analysis :A General Framework Abbas, M; Mandal, S K; Srivastava, M DNA microarrays have emerged as the premier tool for studying gene expression on a genomics scale. They provide a format for the simultaneous measurements of the expression levels of thousands of genes in a single hybridization array. Scientists seeking to harness the potential of this technique are challenged by the large quantities of data produced. For tracking, integrating, qualifying and ultimately deriving scientific insight from the experimental results, various tools are required. In general, a well designed database, an interface for data entry and query, an image analysis software, normalization and filtering routines and software for data analysis and visualization are needed. For data analysis, various statistical and machine learning techniques have been applied. The main task is making groups of genes having the similar expression pattern. For this, methods such as hierarchical clustering, k-means clustering, self organizing maps, neural fretwork, support vector machine etc have been applied.

2006-01-01T00:00:00Z Srivastava, Mukesh Mandal, S K Abbas, M http://dkr.cdri.res.in:8080/xmlui/handle/123456789/509 2010-02-10T20:32:44Z 2002-01-01T00:00:00Z

Discriminating groups with wald test values Srivastava, Mukesh; Mandal, S K; Abbas, M

2002-01-01T00:00:00Z