In recent years the genomes of various species have been decoded. It is an avalanche of disparate data, as genomes can contain a variety of types of messages. For evolutionists, these messages hold many secrets of evolutionary history. If the species share common ancestors, then the contents of their genomes should help decipher that evolutionary history.
For instance, mobile genetic elements are genome segments that can move about, inserting themselves at various locations within the genome. These insertions, according to evolutionists, are a random affair. After all, evolutionists assume that life is a fluke, and biology is one big kludge.
A few years back evolutionists claimed that retroviruses found in primate genomes proved common descent. The retroviruses, it was said, were the perfect evolutionary sign post. They were assumed to insert randomly into the genome and once inserted to stay put. Therefore, if two cousin species shared a similar pattern of retroviruses, then those retroviruses must have been inherited from a common ancestor. It would be too great of a coincidence for the retroviruses to have independently inserted into the two genomes (notice the circular reasoning).
These assumptions are routinely revised. Retroviruses patterns were found that cannot be explained by common descent (e.g., same pattern in only two of three cousin species). Apparently the retroviruses were not such perfect evolutionary sign posts as had been assumed. But evolutionists viewed such cases as anomalies, and rationalized them using ad hoc explanations. In fact I recently received a challenge to find a retrovirus that violates the evolutionary expectation. That's easy. Such falsifiers have long since been discovered. They are just not advertised.
This dynamic has repeated itself with other types of genetic messages, such as pseudogenes and interspersed elements. Here is what Carroll had to say about the latter:
These landmarks are produced by accidental insertions of junk DNA sequences near genes. ... Once [an interspersed element] is inserted, there is no active mechanism for removing it. The insertion of these elements marks a gene in a species, and is then inherited by all species descended from it. They are really perfect tracers of genealogy. 
Except, that is, when they aren't. Like retroviruses, and pseudogenes, interspersed elements occasionally violate the evolutionary pattern. Apparently they are not quite such "perfect tracers of genealogy." To be sure, such outliers are unusual, but if they can be explained by mechanism (rather than inheritance), then so can the others. Carroll concludes that the interspersed elements:
can be explained only by the species sharing a common ancestor. ... biologists have sufficient forensic evidence to determine species' kinship beyond any doubt. 
This is, of course, false. In fact, interspersed elements patterns are explained without resorting to a common ancestor and the tremendous problems with such an explanation.