NEWS

---

flowClust 3.21.2

• Fixed some tests and rewrote the vignette with Rmarkdown.

flowClust 1.8

SIGNIFICANT USER-VISIBLE CHANGES

• flowClust is now Bayesian. You can specify a prior for the means and covariances of each component.

NEW FEATURES

• FlowClust is now bayesian. The user can specify prior information for cell population means and covariances. This is particularly useful for identifying rare cell populations.
• We now check if the covariance matrices can be inverted and if not we stop the algorithm and output NaN for the log likelihood. As a consequence, the function does not stop and can finish all computable values of K.
• Support for the parallel package. The number of cores will automatically be detected. The parallel package needs to be loaded for that.

BUG FIXES