Mutations are the ultimate source of genetic variation upon which natural selection acts 
. As such, mutations play a central role in the evolutionary process. How often new mutations arise has been difficult to determine until recently 
; mostly because mutations are very rare events 
. In addition, many mutations have deleterious fitness effects 
, causing them to be quickly removed by natural selection. Therefore, unless methods are used to minimize selection, deleterious mutations can be undercounted. Estimates based on comparative approaches are further hampered by unknown times of divergence and unknown selection pressures imposed by environmental variation during divergence. Until recently, direct estimation of mutation rates was mostly limited to the analyses of a few genes based on phenotypic assays 
A new and promising approach to studying mutation rates is the combination of mutation accumulation (MA) experiments with whole-genome sequencing using high-throughput technologies 
. The advent of cost-effective sequencing now makes it possible to detect mutations such as substitutions, deletions, insertions, and gene duplications directly at the molecular level in both coding and non-coding regions of the genome 
. MA experiments have a number of advantages over other methods for studying mutation rates. These types of experiments allow spontaneous mutations to accumulate regardless of their effects on fitness, as long as they are not severely deleterious. Natural selection can be relaxed by repeatedly reducing the population size to one individual in asexually reproducing organisms, or to a few closely related individuals (often siblings) in sexually reproducing organisms. This process prevents deleterious (but not fatal) mutations from being eliminated by competition and allows them to be as likely to be fixed by drift as other alleles. Replicated populations sharing a single common ancestor can be propagated under identical experimental conditions for a known number of generations and allowed to accumulate independent, random mutations. These results can be compared among species, with the advantage of matching methodologies across very different life cycles.
Several estimates of eukaryotic, spontaneous, nuclear mutation rates obtained from the whole-genome sequencing of MA lines have been published. Of these, the estimated mutation rate of Arabidopsis thaliana
is the highest. The lowest rate was estimated in Saccharomyces cerevisiae
for asexual haploid cells 
for asexual diploid cells dividing mitotically, and 3.9×10−10
for diploid cells with recombination 
. Drosophila melanogaster
and Caenorhabditis elegans
are estimated to have intermediate mutation rates.
Current mutation rate estimates for the mitochondrial genome from MA experiments suggest a relatively constant mutation rate across different organisms. The highest mitochondrial mutation rate has been estimated in Daphnia pulex
for sexual lines and 1.73×10−7
for asexual lines 
, while the lowest mitochondrial mutation rate was estimated for haploid asexually reproducing yeast at 1.29×10−8
. The estimated mitochondrial mutation rates for C. elegans
and D. melanogaster
are intermediate at 9.7×10−8
, respectively 
We combined an MA experiment and whole-genome sequencing to estimate the spontaneous single nucleotide mutation rate in the social amoeba Dictyostelium discoideum.
This haploid eukaryote is a model system for the evolution of sociality 
, multicellularity, developmental and cellular biology 
, and pathogenicity 
. The social amoeba has a complex life cycle with a vegetative unicellular stage, a social multicellular stage, and a social sexual stage (). During the vegetative stage, single cells live in soil and prey on microorganisms. Upon starvation, single cells start to aggregate, and then, depending on environmental conditions, enter either the social or the sexual stage. During the social, multicellular stage, single cells aggregate and undergo complex behavior that culminates in the formation of a fruiting body, which consists of fertile spores contained within a sorus held aloft by a stalk made of dead cells 
. During fruiting body formation, about 20% of the cells die to form the stalk. Unlike other multicellular organisms that go through a single-cell bottleneck where all cells are essentially clones of the initial zygote, the aggregating cells in D. discoideum
can be genetically different. This can lead to conflict over spore and stalk allocation 
. An alternative to the social cycle is the sexual cycle. During the sexual cycle, two cells of different mating types fuse and form a giant cell that cannibalizes all other aggregating cells, forming a macrocyst that can become dormant and survive harsh environmental conditions 
. This complex life cycle makes it difficult to clearly specify generations, and we have previously called one vegetative cycle combined with fruiting-body formation a social generation 
. During our MA experiment, there is no fruiting stage or sexual stage; the cells remained in the single-cell stage. Therefore, we refer to a single replication event that manifests itself in the division of a single cell into two daughter cells as one generation.
Life cycle of D. discoideum showing the vegetative, social, and sexual.
The haploid nuclear genome of D. discoideum
is 34 MB and contains six chromosomes 
. As of September 11, 2012, 12,646 protein-coding genes have been identified, and 188 pseudogenes have been annotated on dictybase (http://www.dictybase.org
). The genome is very AT-rich (78%), and contains over 11% simple-sequence repeats 
. Roughly 2/3 of the genome is coding sequence. The mitochondrial genome contains 55,564 bases, is slightly less AT-rich than the nuclear genome (72.57%), and encodes 41 genes. On average, one D. discoideum
cell contains about 200 copies of the mitochondrial genome 
Several lines of evidence suggest a relatively low mutation rate for D. discoideum
. Analysis of genetic variation in wild populations indicated low levels of genetic variation, which could be explained by low mutation rates 
. Similarly, D. discoideum
is very resistant to DNA-damaging agents, which suggests efficient DNA-repair mechanisms 
. The most direct evidence for low mutation rates comes from the analysis of microsatellite mutations in the same 90 MA lines used here; D. discoideum
has the lowest per-generation per-repeat mutation rate reported 
We randomly selected three of the same MA lines used to estimate microsatellite mutations (MA31, MA47, and MA55) and sequenced their whole genomes to determine if the genome-wide point mutation rate is also low; or alternatively, if the low mutation rate is a characteristic of the microsatellite regions only.