Selected Research Contributions:

• Application of Boolean implication network to discover stem and progenitor cells in normal and cancer tissue.
  Algorithm maps cancer trajectory Jan 23, 2012 02:04 am | Sarah Moore
  Researchers at the Stanford University School of Medicine have developed a mathematical algorithm to help predict the severity of bladder cancer.
  Computer algorithm used to identify bladder cancer marker , 19 January 2012
  
  

  Researchers have used an innovative mathematical technique to find markers that effectively predict how deadly a cancer -- in this case, bladder cancer -- will be. Full story , Icon
  MedExpress
  Mathematical algorithm used to detect bladder cancer marker: Research
  Advanced technologies give high-resolution snapshot of cancer tissues , 17 November 2011
  
  

  Stanford researchers have melded tools and technologies from engineering, computer science and stem cell biology to analyze hundreds of individual cancer cells and draw the most accurate portrait yet of the cellular composition of human colon cancer tissues. Full story
  
  
• Proposed a new method of predicting developmentally regulated genes using Boolean implications [Sahoo et al. PNAS 2010]. We predicted 62 B cell associated genes where 26 genes show specific defects in B cell function and differentiation from the available 41 mice knockouts.
  In News
  Times of India in Orissa edition on Page 2 , 8 December 2010
  Computational feat helps find genes , 15 March 2010
  
  

  Computer scientist Debashis Sahoo, PhD, has devised a way to take two known genes and then, in a fraction of a second, find new genes involved in the same developmental pathway. Full story
  
  Boolean Logic Unlocks The Key To Finding New Genes in Milliseconds , 10 April 2010
  
  

  A little bit of computer science has made it possible to reduce the time it takes to find new genes from years to milliseconds. Researchers at Stanford University have developed a simple processing method for searching through large public databases of genes and their associated proteins. Using Boolean logic (if A then B, if not C then D, etc), they have found a way to suggest which genes are responsible for different stages of complex chemical processes in our cells. Biologists generally take months or even years to find these genes experimentally, but the Boolean search method takes just a fraction of a second. When widely applied, the new method may accelerate genetic research enormously. Great things happen when sciences come together. Full story
  
  iSGTW: Feature - Stem cell research goes Boolean , 21 April 2010
  

  By looking for asymmetrical relationships, a computer scientist was able to create an algorithm that could cut a multi-year process down to seconds. Full story
• Proposed a new method of analyzing gene expression relationships using Boolean implication network from large gene expression datasets [Sahoo et al. Genome Biology 2008]. We demonstrated that asymmetric Boolean relationships can reveal a wide range of underlying biological phenomena including gender difference, tissue difference, developmental stage and state of differentiation, which cannot be readily identified using popular correlation-based networks. Analysis of high-throughput gene expression datasets is being studied extensively in both academia and industry. Such techniques are also being applied to large healthcare data sets.
• Proposed a new method of discovering step-like changes in gene expression in time-course microarray data sets [Sahoo et al. NAR 2007]. For the first time, we demonstrated that temporal gene expression patterns could be categorized into early and late changes automatically with justifiable statistics. Our software tool is being used in multiple research groups at Stanford and throughout the world. Several papers are currently under preparation at Stanford that demonstrate the usefulness of this tool in understanding biological processes.
• For the first time, we showed that exploring the Boolean state space of large state transition systems could be efficiently parallelized using multithreading [Sahoo et al. DAC 2005]. Our tool is being used largely in Fujitsu labs. Analysis of large Boolean state space is an extensively studied area of research and heavily used in verification of large digital systems.