I like to think of inbreeding (line-breeding) kinda like a compression algorithm (specifically lossy compression / irreversible compression). You start with a file (dog population), remove data from it (genes), while still producing a good estimate of the original.

Genes are a lot like bits of data – individually they don’t mean much, but together they form a dog – a collection of data (a dog) completes a file (a dog population).

In a well designed lossy compression, the end result will be an acceptable approximation of the original, but with a smaller file size. Size can be reduced significantly before degradation is noticed.

Recompressing a file that has already been compressed via lossy compression will result in further degradation and smaller file size. Repeating this practice will eventually results in an exponential increase in defects and an exponential decline in approximation (accuracy).

In a well designed line breeding program, the output will also result in an acceptable approximation of the original, while reducing the population size considerably (due to diversity being discarded). Population size can be reduced significantly before degradation is noticed.

Continuing to inbreed on inbred dogs results in further degradation and smaller population size. Repeating this practice will eventually results in an exponential increase in defects and an exponential decline in approximation (accuracy).

Tagged with: