$\bf{Context}$. We investigate the validity of the claim that invokes two extreme exoplanetary system candidates around the pulsating B-type subdwarfs KIC 10001893 and KIC 5807616 from the primary $\it{Kepler}$ field. $\bf{Aims}$. Our goal was to find characteristics and the source of weak signals that are observed in these subdwarf light curves. $\bf{Methods}$. To achieve this, we analyzed short- and long-cadence $\it{Kepler}$ data of the two stars by means of a Fourier transform and compared the results to Fourier transforms of simulated light curves to which we added exoplanetary signals. The long-cadence data of KIC 10001893 were extracted from CCD images of a nearby star, KIC 10001898, using a point spread function reduction technique. $\bf{Results}$. It appears that the amplitudes of the Fourier transform signals that were found in the low-frequency region depend on the methods that are used to extract and prepare $\it{Kepler}$ data. We demonstrate that using a comparison star for space telescope data can significantly reduce artifacts. Our simulations also show that a weak signal of constant amplitude and frequency, added to a stellar light curve, conserves its frequency in Fourier transform amplitude spectra to within 0.03 $\mu$Hz. $\bf{Conclusions}$. Based on our simulations, we conclude that the two low-frequency Fourier transform signals found in KIC 5807616 are likely the combined frequencies of the lower amplitude pulsating modes of the star. In the case of KIC 10001893, the signal amplitudes that are visible in the light curve depend on the data set and reduction methods. The strongest signal decreases significantly in amplitude when KIC 10001898 is used as a comparison star. Finally, we recommend that the signal detection threshold is increased to 5 $\sigma$ (or higher) for a Fourier transform analysis of $\it{Kepler}$ data in low-frequency regions.