Joint Analysis Of Multi-omics Data Of HCC Cell Lines With Stepwise Increasing Metastasis Ability

With the completion of the sequencing of all genes and structural elements on human chromosome 1 in mid-2016, the dawn of post-genomic era came. When decoding the inner knowledges of human genome, the so-called “the book of life”, researches kept developing and revolutionizing the relevant technologies. Omics thrives under these circumstances and quality became a hotspot of biological researches. Among kinds of “-omics” technologies, high-throughput sequencing based transcriptomics and tandem mass spectrometry based proteomics have the advantages of high coverage, high precision and high throughput, making them the most important technologies in the studies of-omics.The HTS and MS/MS have been progressed vastly recently and both transcriptomics and proteomics have been applied in various biological researches. But few of these researches paid equal attention to both of them. Based on central dogma, genetic information in one gene is passed to protein product through its transcript, so there exist a direct correspondence between m RNA transcripts and generated protein products. However, many studies showed poor correlation between transcriptome and proteome data from same cell under similar conditions. These reveal complex regulations in and after the process of transcription and translation. To make a joint analysis of these tow data sources, one can find unique viewpoint to get the insight of system biology.The researches involved in joint analysis of omics data have been polarized. On one hand bioinformaticians have developed a bunch of algorithms and strategies using statistics and computer science methods which lack of practical application. On the other hand, most researches in omics area who did not expertise in data analysis, which led to primitively analysis of multi-omics data. Under these circumstances, we present this joint analysis of transcriptome and proteome of HCC cell lines of invasion and metastasis model. The details of this research are as follow.1. Omics data have inevitable systematic false positive error. To eliminate these errors to maximum degree, the original data must be properly processed and the data quality must be strictly controlled. This study based on RNA-Seq generated transcriptome data and SILAC labeled quantitative proteome data. After choosing pertinent data analysis workflow and quality control, we gained access to high- quality multi-omics profiling. Using Edge R and ANOVA, we screened differentially expressed genes / proteins from transcriptome and proteome data, respectively.2. Based on DEG / DEP profiling, we analyzed the correlations between bio samples and omics data. By defining accumulated difference, we filtered 530 genes with most differently expressed between transcriptome and proteome, then analyzed if they share functional similarity. In order to draw an unbiased picture of dynamic proteome change among cell lines with increasing metastasis ability, we clustered the 1,273 DEP into 4 clusters, then confirmed certain feature in different cluster functionally.3. Under the guidance of multi-omics data analysis workflow, we analyzed the big data generated by the Chromosome-Centric Human Proteome Project. By jointly analyzing this dataset, which contained genome, transcriptome, translatome and proteome data of HCC cell lines, we demonstrated the genetic information passing through the multi-omics data. Based these results, we characterized the missing protein using regular bio samples by systematic enrichment strategies.Above all, this study built a complete workflow of multi-omics data based on transcriptome and proteome profiling of invasion and metastasis model of HCC cell lines, including original data processing, data quality control, DEG / DEP screening, omics data functional clustering and protein-protein interaction network construction. Then we applied this workflow into integrated multi-omics data analysis such as CHPP, which threw light on the further investigation of missing protein.