Email updates

Keep up to date with the latest news and content from BMC Genomics and BioMed Central.

Open Access Research article

Translational signatures and mRNA levels are highly correlated in human stably expressed genes

Sergio R P Line12*, Xiaoming Liu3, Ana Paula de Souza1 and Fuli Yu2

Author Affiliations

1 Piracicaba Dental School, University of Campinas, PO Box 52, Piracicaba, SP, 13414-903, Brazil

2 The Human Genome Sequencing Center, Baylor College of Medicine, Houston, TX, 77030, USA

3 Human Genetics Center, School of Public Health, The University of Texas Health Science Center at Houston, Houston, TX, 77030, USA

For all author emails, please log on.

BMC Genomics 2013, 14:268  doi:10.1186/1471-2164-14-268

Published: 19 April 2013



Gene expression is one of the most relevant biological processes of living cells. Due to the relative small population sizes, it is predicted that human gene sequences are not strongly influenced by selection towards expression efficiency. One of the major problems in estimating to what extent gene characteristics can be selected to maximize expression efficiency is the wide variation that exists in RNA and protein levels among physiological states and different tissues. Analyses of datasets of stably expressed genes (i.e. with consistent expression between physiological states and tissues) would provide more accurate and reliable measurements of associations between variations of a specific gene characteristic and expression, and how distinct gene features work to optimize gene expression.


Using a dataset of human genes with consistent expression between physiological states we selected gene sequence signatures related to translation that can predict about 42% of mRNA variation. The prediction can be increased to 51% when selecting genes that are stably expressed in more than 1 tissue. These genes are enriched for translation and ribosome biosynthesis processes and have higher translation efficiency scores, smaller coding sequences and 3 UTR sizes and lower folding energies when compared to other datasets. Additionally, the amino acid frequencies weighted by expression showed higher correlations with isoacceptor tRNA gene copy number, and smaller absolute correlation values with biosynthetic costs.


Our results indicate that human gene sequence characteristics related to transcription and translation processes can co-evolve in an integrated manner in order to optimize gene expression.