Gene signatures identified using Computational biology approaches to distinguish between former smokers, current smokers, and non-smokers.
By Ratneshwar Thakur Published in The Hawk
The new study, published in “Computational Toxicology”, reveals how smoke exposure influences the gene expression levels- from mice to human. These gene signatures may give us a really good clue to understanding what might have happened due to smoke exposure.
The authors of this article were part of three teams that participated in a world-wide computational challenge organized by Philip Morris International (PMI), Switzerland and were ranked at the top in this crowdsourcing initiative (https://sbvimprover.com/challenge-4) that was designed to determine whether dysregulation of blood expression of a common set of genes can classify both human and mouse subjects into cigarette smoke exposure groups.
“The common motivation to participate in the challenge was to assess our approaches to develop accurate prediction models based on high-dimensional datasets, and also to understand how translatable the response to toxicants is between mouse and human”, said the Investigators.
The ability to translate the impact of such toxicants from animals to humans is key in systems toxicology, which is enabled by the ability to profile tens of thousands of molecules (e.g. mRNAs) in biological samples.
In this challenge, participants applied their prediction models to a blinded test dataset which was obtained by expression profiling of blood samples from 27 smokers, 26 former smokers, and 28 never-smokers. The mouse training dataset was based on a 7-month cigarette smoke inhalation study conducted with mice and included three groups of animals: exposed to smoke for 7 months (equivalent to a human current smoker), exposed to smoke for 2 months followed by exposure to air (equivalent to a human former smoker) and mice continuously exposed to air (equivalent to human never-smoker).
“This study is interesting because it shows that a common gene signature predicts with good accuracy current exposure to cigarette smoke in both human and mouse, and that the microarray technology and our computational methods reached the maturity needed to obtain reliable results in other similar research settings’’, says Adi L. Tarca, Associate Professor, Wayne State University, School of Medicine.
“The publication has reported several genes that are different in the former smoker and current smoker. Additionally, it opens the industries to seek crowdsourcing solutions to their research problems”, says Dr. Sandeep Kumar Dhanda, Bioinformatics postdoc, La Jolla Institute for Allergy and Immunology, La Jolla, California, USA
“Prediction of outcomes from high-dimensional data is a research topic with applications in many areas of biomedicine, as researchers use these types of data to develop models to predict future onset of disease or patient response to treatments. Moreover, for those working in toxicology, the data presented in this study are useful to assess the suitability of the mouse as a model organism to predict the response to toxicants in human based on observation in the mouse model”, says Prof. Tarca.
“We observed that smoking affects the gene expression profile of an individual, but this effect is minimized if a person quit smoking for more than the year. These gene expression profiles can be used as a platform to decode the underlying changes going on in the body of a smoker”, says post-doctoral researchers Dr. Sandeep K. Dhanda and Dr. Rahul Kumar.