Professor Yutaka SUZUKI【Key Words】genome, transcriptome, epigenome, cancer genome, genomic drug discovery
It is still remaining mostly unknown which of the variations or mutations occurring in the human genomes contribute to etiology of diseases. We employ versatile applications of next generation sequencing technologies, such as Whole Genome/Exome Seq, RNA Seq, ChIP Seq and Bisulfite Seq to understand the biological meaning of the identified genomic mutations.
As collaboration with several hospitals and laboratories of clinical sequencing,we have analyzed the mutation patterns of various types of cancers, including lung, colon and stomach cancers. We have found that the mutated genes are mostly distinct depending on patients and cancer types, with rare exceptions of the TP53, KRAS and EGFR genes. With rare mutual overlaps, it is difficult to statistically discriminate so-called driver mutations, which serve as a direct driving force to carcinogenesis, from so-called passenger mutations, which occur in the human genomes as a consequence of chromosomal instability in cancers, thus, have no functional relevance. Moreover,in spite of supposed importance, almost no clue has been obtained for the mutations which invoke abnormal transcriptional regulations. To address these issues, we have established an experimental system to collect genome,epigenome and transcriptome data from the same cellular material and have started the data production. By integrating such multi-omics data, we are investigating epigenomic and transcriptomic consequences of the genomic mutations.
Recent genome-wide analyses have revealed that gene expression regulations, such as the regulations at transcriptional elongation, RNA logistic and RNA degradation, play no less important roles than transcriptional initiations. We are trying to develop a new method to evaluate the contribution from these factors, using the latest genomerelated technologies. We have constructed an experimental system in which correlation between DNA mutations at every base position can be associated with promoter activities for thousands of genes simultaneously. Generated data is further processed to construct a model, using machinelearning and statistical inference technologies, to predict eventual transcript levels. We are also including the data obtained from the emerging technologies measuring posttranscriptional regulatory factors to the model. Eventually, we believe such a model should be essential to understand biological meaning of the genomic variations of regulatory roles in the humans.
Frequently, behaviors of human immune systems in responds to pathogens are significantly different in field from those in laboratory conditions. We have a field base in Indonesia and are analyzing the mutual correlation between the host-pathogens at every omics layer,particularly focusing on malaria parasites.
Suzuki et al. Nucleic Acids Research. 2015
Irie et al. Nucleic Acids Research. 2011
Yamagishi et al. Genome Research. 2014
