Abstract
BACKGROUND: Gene expression analysis by reverse transcription quantitative PCR (qPCR) is the most widely used method for analyzing the expression of a moderate number of genes and also for the validation of microarray results. Several issues are crucial for a successful qPCR study, particularly the selection of internal reference genes for normalization and efficiency determination. There is no agreement on which method is the best to detect the most stable genes neither on how to perform efficiency determination. In this study we offer a comprehensive evaluation of the characteristics of reference gene selection methods and how to decide which one is more reliable when they show discordant outcomes. Also, we analyze the current efficiency calculation controversy. Our dataset is composed by gonad samples of turbot at different development times reared at different temperatures. Turbot (Scophthalmus maximus) is a relevant marine aquaculture European species with increasing production in the incoming years. Since females largely outgrow males, identification of genes related to sex determination, gonad development and reproductive behavior, and analysis of their expression profiles are of primary importance for turbot industry. RESULTS: We analyzed gene stability of six reference genes: RPS4, RPL17, GAPDH, ACTB, UBQ and B2M using the comparative delta-CT method, Bestkeeper, NormFinder and GeNorm approaches in gonad samples of turbot. Supported by descriptive statistics, we found NormFinder to be the best method, while on the other side, GeNorm results proved to be unreliable. According to our analysis, UBQ and RPS4 were the most stable genes, while B2M was the least stable gene. We also analyzed the efficiency calculation softwares LinRegPCR, LREanalyzer, DART and PCR-Miner and we recommend LinRegPCR for research purposes since it does not systematically overestimate efficiency. CONCLUSION: Our results indicate that NormFinder and LinRegPCR are the best approaches for reference gene selection and efficiency determination, respectively. We also recommend the use of UBQ and RPS4 for normalization of gonad development samples in turbot.