Banana streak virus (BSV) is a badnavirus infecting Musa species and inducing Banana streak disease, affecting plant growth and fruit production. BSV exists under two forms: episomal and/or integrated in host genome (referred as “endogenous BSV”, eBSV). Among these viral integrations, four BSV species (BSGFV, BSIMV, BSOLV, and BSMYV) are known to induce eBSV-derived infections under biotic and abiotic stresses.My PhD project aims to follow eBSV-derived infections kinetics in interspecific Musa hybrids (hosting eBSV sequences), over an entire crop cycle. I used qPCR as a quantitative tool to compare BSV infection between different samples, and across time. Because sequence similarity between eBSV and BSV overcomes 99%, I used relative quantification between infected and healthy samples of the same banana genotype hosting eBSV. I refer the healthy, viral particles-free samples as “calibrator”. Viral DNA amplification in the calibrator was supposed to be constant, since eBSV DNA matrix is stable. However, I observed significant variability in eBSV-DNA amplification among a large population of calibrators. This might result from transitory eBSV-DNA expresssion, not encaspidated and therefore escaping serological diagnostics.This result makes this “calibrator” system unreliable for a diagnostic purpose of low infections in banana plants having viral integrations, as standard IC-PCR (immune-capture) and RCA methods have shown better sensibility. Nevertheless, I still use this protocol in my experiments for relative quantification analysis by including a “buffer area” on both sides of the amplification baseline, reflecting viral DNA amplification variations I observed.