Elongation factor-1α, a putative single-copy nuclear gene, has divergent sets of paralogs in an arachnid.

Identification of paralogy in candidate nuclear loci is an important prerequisite in phylogenetics and statistical phylogeography, but one that is often overlooked. One marker commonly assumed to be a single-copy gene and claimed to harbor great utility for inferring recent divergences is elongation factor-1α (EF-1α). To test this hypothesis, we systematically cloned EF-1α in three disjunct populations of the harvestman Metasiro americanus. Here we show that EF-1α has a large number of paralogs in this species. The paralogs do not evolve in a concerted manner, and the paralogs diverged prior to the population divergence. Moreover, the paralogs of M. americanus are not comparable to the highly divergent EF-1α paralogs found in bees and spiders, which are easily recognized and separated through the use of specific primers. We demonstrate statistically that our detection of paralogs cannot be attributed to amplification error. The presence of EF-1α paralogs in M. americanus prevents its use in statistical phylogeography, and the presence of out-paralogs argues against its use in phylogenetic inference among recently diverged clades. These data contradict the common assumption that EF-1α is for most or all taxa a single-copy gene, or that it has a small number of paralogs that are homogenized through gene conversion, unequal crossing over, or other processes.