The Human Liver Secretome: A Transcriptome-Based Survey

Proteins secreted from cells into the extra-cellular space play a variety of important roles in intercellular communication, and have the potential as targets in disease diognosis and theropeuties. Experimental techniques confront some difficults upon exploring the secretome. Fortunately, increasing data of genome and transcriptomes make it possible to understand the secretome via bioinformatic approaches.Liver is the second largest organ next to brain and has many important functions. Researches on liver secretome will benefit us to uncover the function of liver itself, the relationship between liver and other tissues or organs (e.g. liver and plasma), and the secretion pathway of liver proteins.In this study, based on the transcriptome data, we explored the human liver secretome via bioinformatic approaches.Firstly, we set up a system to predict a secretome, and evaluated the accuration of the system. The results showed that the precision and recalling of the prediction system are 78.8% and 80.7% respectively. In other words, the false positive and false negative of the system are 21.2% and 19.3% respectively. Then, we traced the proteins contributed to the false results, and found that lysosome prteins, membrane proteins and ER proteins are the main sources leading to the false positive results. Multiplly localized proteins and some secreted proteins cause the false negative results.Secondly, based on the transcriptome data, we predicted and defined the human liver secretome. The secretome consists of 1597 proteins encoded by 1037 genes.Thirdly, we characterized the human liver secretome in seven aspects:(1) InterPro domain analysis. Of all the 1597 proteins of human liver secretome, 1414(88.54%) proteins have InterPro domain matches, covering 745 InterPro domains. The top 20 InterPro domains include 4 members of epidermal growth...