How to organize genomic data
A long-term study conducted by employees of several major US universities under the leadership of the deputy director of the Simons Center for Data Analysis, Professor Olga Troyanskaya, provided a fundamentally new level of data by a specialist demonstrating the joint work of genes of various tissues and types of human cells, ensuring the performance of the functions of these tissues both in a healthy body, so it is with various diseases. Moreover, the work done by the authors demonstrated the possibility of joint use of computer calculations and statistical analysis methods for grouping and analyzing very large and extremely diverse genomic data.
The figure shows one of the 144 functional genetic systems identified in the study.
As part of the work, the researchers collected and integrated data obtained during 38,000 genome-wide experiments described in about 14,000 publications. The data arrays used contained data on the functions of DNA and proteins in the cells of both healthy and people suffering from various diseases.
Using integrative computational analysis, the authors first isolated functional genetic relationships for various types of tissues from the entire set of data. After that, by comparing the identified tissue-specific functional signals with the results of genome-wide studies of associations in DNA in various diseases, they identified statistical links between genes and diseases that could not be detected by any other methods.
The resulting technique, called the systematic association study or NetWAS (from the English network-guided association study), provides integration of quantitative genetics and functional genomics. This increases the power of genome-wide association studies and makes it possible to identify genes whose mutations underlie complex human diseases. Since NetWAS is a fully data-driven technique, it eliminates the appearance of errors due to the better knowledge of certain genes and mechanisms, and allows you to identify previously unknown associations.
The result of the work was the identification of 144 functional genetic interactions for various human organs, including kidneys, liver and brain. Moreover, the authors described functional disorders of genes in a variety of diseases, including hypertension, diabetes mellitus and obesity.
An important point is the fact that functional systems of genetic interactions have already been described in animal models. However, for human tissues, this task was impossible without computer processing of huge amounts of data. Many types of human cells involved in the development of various diseases cannot be studied by traditional direct experiments, therefore, the development of a technique that allows processing such data arrays is an extremely important achievement.
The authors note that their results will help not only to understand the normal functioning of genes, but also to develop newer drugs, as well as to optimize the methods of using existing therapeutic approaches. In particular, they will increase the efficiency of identifying target genes for therapy and predicting previously not considered interactions between drugs and their side effects.
The researchers also created an interactive GIANT server (from the English Genome-scale Integrated Analysis of Networks in Tissues – genome-wide integrated analysis of tissue systems). It allows users to study the genetic systems of various tissues, compare them with each other and analyze the data obtained during genetic research to identify the genes that cause various diseases.
Article by Olga G Troyanskaya et al. Understanding multicellular function and disease with human tissue-specific networks is published in the journal Nature Genetics.
Evgeniya Ryabtseva
Portal "Eternal youth" http://vechnayamolodost.ru based on Simons Foundation materials:
Olga Troyanskaya Brings Order to Big Data of Human Biology.
29.04.2015