The shape of the kernel

The shape of the kernel is defined by the kernel function. The shape of the kernel determines how the mRNA signal is smoothed.

We adopt the use of the Gaussian kernel due to it’s popular use in signal processing, however other kernel functions can be used: - we have had success in using semi-circle kernels when applied to ISS data of the human pancreas - the Epanechnikov kernel minimizes AMISE and has therefore been described as optimal

The following exmaples shows how you can apply a semicircular kernel instead of a Gaussian.

# code to change the shape of the kernel (@sebastiantiesmeyer)