Welcome to danielcrouch.net

Uses poisson regression ('Lindsey's method') and expectation maximisation to fit constrained likelihood models to histogram data. Likelihood methods are constrained to enable 'prior splitting', fitting separate mixtures for observed values that have true values lying above and below thresholds. Estimating priority False Discovery Rates (priorityFDRs) and false sign probabilities are two main uses.
Link (on github)
Paper (open access)

An API applying priorsplitteR to estimate the probability that, given ratings for two items, item A is truly rated better on average than item B.
RaterBayes website
On RapidAPI

LEAPFrOG (Likelihood Estimation of Admixture in Parents From Offspring Genotype, an R package)
This program provides maximum likelihood estimates of individuals' admixture proportions, based on their DNA data and DNA data for 'reference' populations (e.g. Africa, Europe, Asia) provided by the user. Furthermore, based only on these data, it provides estimates of the admixture proportions for the individuals' parents, even though their DNA is unobserved. There are two functions, one for use with 'phased' DNA data and one for unphased.
Link (on github)
Paper (open access)

REGENT (Risk Estimation for Genetic and Environmental Traits, an R package)
This provides estimates of relative risks and odds ratios for diseases based on patients' DNA data, and also empirically-estimated confidence intervals for these risk ratios. A risk categorisation is performed (e.g. individuals are assigned low/medium/high risks) based on how these confidence intervals vary and overlap between individuals within a simulated population.
Link (on CRAN)
Paper (open access)
User manual (pdf)