We re-analyse the recently published HARPS and PFS velocities of the nearby K dwarf GJ 221 that have been reported to contain the signatures of two planets orbiting the star. Our goal is to see whether the earlier studies discussing the system fell victims of false negative detections. We perform the analyses by using an independent statistical method based on posterior samplings and model comparisons in the Bayesian framework that is known to be more sensitive to weak signals of low-mass planets. According to our analyses, we find strong evidence in favour of a third candidate planet in the system corresponding to a cold sub-Saturnian planet with an orbital period of 500 days and a minimum mass of 29 $M_{oplus}$. Application of sub-optimal signal detection methods can leave low-amplitude signals undetected in radial velocity time-series. Our results suggest that the estimated statistical properties of low-mass planets can thus be biased because several signals corresponding to low-mass candidate planets may have gone unnoticed. This also suggests that the occurrence rates of such planets based on radial velocity surveys might be underestimated.
Download