Finding the right coverage: the impact of coverage and sequence quality on single nucleotide polymorphism genotyping error rates

Restriction‐enzyme‐based sequencing methods enable the genotyping of thousands of single nucleotide polymorphism (SNP) loci in nonmodel organisms. However, in contrast to traditional genetic markers, genotyping error rates in SNPs derived from restriction‐enzyme‐based methods remain largely unknown. Here, we estimated genotyping error rates in SNPs genotyped with double digest RAD sequencing from Mendelian incompatibilities in known mother–offspring dyads of Hoffman's two‐toed sloth (Choloepus hoffmanni) across a range of coverage and sequence quality criteria, for both reference‐aligned and de novo‐assembled data sets. Genotyping error rates were more sensitive to coverage than sequence quality and low coverage yielded high error rates, particularly in de novo‐assembled data sets. For example, coverage ≥5 yielded median genotyping error rates of ≥0.03 and ≥0.11 in reference‐aligned and de novo‐assembled data sets, respectively. Genotyping error rates declined to ≤0.01 in reference‐aligned data sets with a coverage ≥30, but remained ≥0.04 in the de novo‐assembled data sets. We observed approximately 10‐ and 13‐fold declines in the number of loci sampled in the reference‐aligned and de novo‐assembled data sets when coverage was increased from ≥5 to ≥30 at quality score ≥30, respectively. Finally, we assessed the effects of genotyping coverage on a common population genetic application, parentage assignments, and showed that the proportion of incorrectly assigned maternities was relatively high at low coverage. Overall, our results suggest that the trade‐off between sample size and genotyping error rates be considered prior to building sequencing libraries, reporting genotyping error rates become standard practice, and that effects of genotyping errors on inference be evaluated in restriction‐enzyme‐based SNP studies.     

APIS: An auto‐adaptive parentage inference software that tolerates missing parents

In the context of parentage assignment using genomic markers, key issues are genotyping errors and an absence of parent genotypes because of sampling, traceability or genotyping problems. Most likelihood‐based parentage assignment software programs require a priori estimates of genotyping errors and the proportion of missing parents to set up meaningful assignment decision rules. We present here the R package APIS, which can assign offspring to their parents without any prior information other than the offspring and parental genotypes, and a user‐defined, acceptable error rate among assigned offspring. Assignment decision rules use the distributions of average Mendelian transmission probabilities, which enable estimates of the proportion of offspring with missing parental genotypes. APIS has been compared to other software (CERVUS, VITASSIGN), on a real European seabass (Dicentrarchus labrax) single nucleotide polymorphism data set. The type I error rate (false positives) was lower with APIS than with other software, especially when parental genotypes were missing, but the true positive rate was also lower, except when the theoretical exclusion power reached 0.99999. In general, APIS provided assignments that satisfied the user‐set acceptable error rate of 1% or 5%, even when tested on simulated data with high genotyping error rates (1% or 3%) and up to 50% missing sires. Because it uses the observed distribution of Mendelian transmission probabilities, APIS is best suited to assigning parentage when numerous offspring (>200) are genotyped. We have demonstrated that APIS is an easy‐to‐use and reliable software for parentage assignment, even when up to 50% of sires are missing.     

Effects of genotyping errors on parentage exclusion analysis

Genetic markers are widely used to determine the parentage of individuals in studies of mating systems, reproductive success, dispersals, quantitative genetic parameters and in the management of conservation populations. These markers are, however, imperfect for parentage analyses because of the presence of genotyping errors and undetectable alleles, which may cause incompatible genotypes (mismatches) between parents and offspring and thus result in false exclusions of true parentage. Highly polymorphic markers widely used in parentage analyses, such as microsatellites, are especially prone to genotyping errors. In this investigation, I derived the probabilities of excluding a random (related) individual from parentage and the probabilities of Mendelian-inconsistent errors (mismatches) and Mendelian-consistent errors (which do not cause mismatches) in parent-offspring dyads, when a marker having null alleles, allelic dropouts and false alleles is used in a parentage analysis. These probabilities are useful in evaluating the impact of various types of genotyping errors on the information content of a set of markers in and thus the power of a parentage analysis, in determining the threshold number of genetic mismatches that is appropriate for a parentage exclusion analysis and in estimating the rates of genotyping errors and frequencies of null alleles from observed mismatches between known parent-offspring dyads. These applications are demonstrated by numerical examples using both hypothetical and empirical data sets and discussed in the context of practical parentage exclusion analyses.     

Microsatellite genotyping errors: detection approaches, common sources and consequences for paternal exclusion

Microsatellite genotyping errors will be present in all but the smallest data sets and have the potential to undermine the conclusions of most downstream analyses. Despite this, little rigorous effort has been made to quantify the size of the problem and to identify the commonest sources of error. Here, we use a large data set comprising almost 2000 Antarctic fur seals Arctocephalus gazella genotyped at nine hypervariable microsatellite loci to explore error detection methods, common sources of error and the consequences of errors on paternal exclusion. We found good concordance among a range of contrasting approaches to error-rate estimation, our range being 0.0013 to 0.0074 per single locus PCR (polymerase chain reaction). The best approach probably involves blind repeat-genotyping, but this is also the most labour-intensive. We show that several other approaches are also effective at detecting errors, although the most convenient alternative, namely mother-offspring comparisons, yielded the lowest estimate of the error rate. In total, we found 75 errors, emphasizing their ubiquitous presence. The most common errors involved the misinterpretation of allele banding patterns (n = 60, 80%) and of these, over a third (n = 22, 36.7%) were due to confusion between homozygote and adjacent allele heterozygote genotypes. A specific test for whether a data set contains the expected number of adjacent allele heterozygotes could provide a useful tool with which workers can assess the likely size of the problem. Error rates are also positively correlated with both locus polymorphism and product size, again indicating aspects where extra effort at error reduction should be directed. Finally, we conducted simulations to explore the potential impact of genotyping errors on paternity exclusion. Error rates as low as 0.01 per allele resulted in a rate of false paternity exclusion exceeding 20%. Errors also led to reduced estimates of male reproductive skew and increases in the numbers of pups that matched more than one candidate male. Because even modest error rates can be strongly influential, we recommend that error rates should be routinely published and that researchers make an attempt to calculate how robust their analyses are to errors.     

DNA-based parentage verification in two Australian goat herds

This study aimed to evaluate a set of DNA markers for their effectiveness in parentage inference, to quantify the level of pedigree errors in Australian Angora and Cashmere goat herds using different pedigree recording methods, and to investigate genotype mismatches between parent and offspring. The 14 microsatellite markers evaluated in this study provided a high level of power (probability of exclusion, PE >99.70%) for parentage testing. The extent of PE depended on polymorphic information content (PIC) and number of alleles for each marker. The minimum number of MS markers essential for accurate determination of parentage was 12, when neither parent is known (PE1) and 10, when one parent is known (PE2). In both populations, the error rates of recorded sire and dam pedigree were significant, averaging around 12%. The error rates of sire and dam pedigree varied considerably between the two populations, reflecting management differences on the two properties. Of 14 MS markers, one locus, SRCRSP07, had null alleles present in the heterozygous state. This null allele was revealed by mismatches of genotypes of parent-offspring pairs. Highly significant deviation from Hardy–Weinberg Equilibrium and significant heterozygote deficiency was also observed at this locus.     

Noninvasive genotyping and Mendelian analysis of microsatellites in African savannah elephants

We obtained fresh dung samples from 202 (133 mother-offspring pairs) savannah elephants (Loxodonta africana) in Samburu, Kenya, and genotyped them at 20 microsatellite loci to assess genotyping success and errors. A total of 98.6% consensus genotypes was successfully obtained, with allelic dropout and false allele rates at 1.6% (n = 46) and 0.9% (n = 37) of heterozygous and total consensus genotypes, respectively, and an overall genotyping error rate of 2.5% based on repeat typing. Mendelian analysis revealed consistent inheritance in all but 38 allelic pairs from mother-offspring, giving an average mismatch error rate of 2.06%, a possible result of null alleles, mutations, genotyping errors, or inaccuracy in maternity assignment. We detected no evidence for large allele dropout, stuttering, or scoring error in the dataset and significant Hardy-Weinberg deviations at only two loci due to heterozygosity deficiency. Across loci, null allele frequencies were low (range: 0.000-0.042) and below the 0.20 threshold that would significantly bias individual-based studies. The high genotyping success and low errors observed in this study demonstrate reliability of the method employed and underscore the application of simple pedigrees in noninvasive studies. Since none of the sires were included in this study, the error rates presented are just estimates.     

Importance of a pilot study for non-invasive genetic sampling: genotyping errors and population size estimation in red deer

Population size information is critical for managing endangered or harvested populations. Population size can now be estimated from non-invasive genetic sampling. However, pitfalls remain such as genotyping errors (allele dropout and false alleles at microsatellite loci). To evaluate the feasibility of non-invasive sampling (e.g., for population size estimation), a pilot study is required. Here, we present a pilot study consisting of (i) a genetic step to test loci amplification and to estimate allele frequencies and genotyping error rates when using faecal DNA, and (ii) a simulation step to quantify and minimise the effects of errors on estimates of population size. The pilot study was conducted on a population of red deer in a fenced natural area of 5440 ha, in France. Twelve microsatellite loci were tested for amplification and genotyping errors. The genotyping error rates for microsatellite loci were 0–0.83 (mean=0.2) for allele dropout rates and 0–0.14 (mean=0.02) for false allele rates, comparable to rates encountered in other non-invasive studies. Simulation results suggest we must conduct 6 PCR amplifications per sample (per locus) to achieve approximately 97% correct genotypes. The 3% error rate appears to have little influence on the accuracy and precision of population size estimation. This paper illustrates the importance of conducting a pilot study (including genotyping and simulations) when using non-invasive sampling to study threatened or managed populations.     

Using blocks of linked single nucleotide polymorphisms as highly polymorphic genetic markers for parentage analysis

Single nucleotide polymorphisms (SNPs) are plentiful in most genomes and amenable to high throughput genotyping, but they are not yet popular for parentage or paternity analysis. The markers are bi-allelic, so individually they contain little information about parentage, and in nonmodel organisms the process of identifying large numbers of unlinked SNPs can be daunting. We explore the possibility of using blocks of between three and 26 linked SNPs as highly polymorphic molecular markers for reconstructing male genotypes in polyandrous organisms with moderate (five offspring) to large (25 offspring) clutches of offspring. Haplotypes are inferred for each block of linked SNPs using the programs Haplore and Phase 2.1. Each multi-SNP haplotype is then treated as a separate allele, producing a highly polymorphic, 'microsatellite-like' marker. A simulation study is performed using haplotype frequencies derived from empirical data sets from Drosophila melanogaster and Mus musculus populations. We find that the markers produced are competitive with microsatellite loci in terms of single parent exclusion probabilities, particularly when using six or more linked SNPs to form a haplotype. These markers contain only modest rates of missing data and genotyping or phasing errors and thus should be seriously considered as molecular markers for parentage analysis, particularly when the study is interested in the functional significance of polymorphisms across the genome.     

Genotyping‐free parentage assignment using RAD‐seq reads

Parentage assignment is defined as the identification of the true parents of one focal offspring among a list of candidates and has been commonly used in zoological, ecological, and agricultural studies. Although likelihood‐based parentage assignment is the preferred method in most cases, it requires genotyping a predefined set of DNA markers and providing their population allele frequencies. In the present study, we proposed an alternative method of parentage assignment that does not depend on genotype data and prior information of allele frequencies. Our method employs the restriction site‐associated DNA sequencing (RAD‐seq) reads for clustering into the overlapped RAD loci among the compared individuals, following which the likelihood ratio of parentage assignment could be directly calculated using two parameters—the genome heterozygosity and error rate of sequencing reads. This method was validated on one simulated and two real data sets with the accurate assignment of true parents to focal offspring. However, our method could not provide a statistical confidence to conclude that the first ranked candidate is a true parent.     

DNA-based parentage verification in two Australian goat herds

This study aimed to evaluate a set of DNA markers for their effectiveness in parentage inference, to quantify the level of pedigree errors in Australian Angora and Cashmere goat herds using different pedigree recording methods, and to investigate genotype mismatches between parent and offspring. The 14 microsatellite markers evaluated in this study provided a high level of power (probability of exclusion, PE >99.70%) for parentage testing. The extent of PE depended on polymorphic information content (PIC) and number of alleles for each marker. The minimum number of MS markers essential for accurate determination of parentage was 12, when neither parent is known (PE1) and 10, when one parent is known (PE2). In both populations, the error rates of recorded sire and dam pedigree were significant, averaging around 12%. The error rates of sire and dam pedigree varied considerably between the two populations, reflecting management differences on the two properties. Of 14 MS markers, one locus, SRCRSP07, had null alleles present in the heterozygous state. This null allele was revealed by mismatches of genotypes of parent-offspring pairs. Highly significant deviation from Hardy–Weinberg Equilibrium and significant heterozygote deficiency was also observed at this locus.     

megasat: automated inference of microsatellite genotypes from sequence data

megasat is software that enables genotyping of microsatellite loci using next‐generation sequencing data. Microsatellites are amplified in large multiplexes, and then sequenced in pooled amplicons. megasat reads sequence files and automatically scores microsatellite genotypes. It uses fuzzy matches to allow for sequencing errors and applies decision rules to account for amplification artefacts, including nontarget amplification products, replication slippage during PCR (amplification stutter) and differential amplification of alleles. An important feature of megasat is the generation of histograms of the length–frequency distributions of amplification products for each locus and each individual. These histograms, analogous to electropherograms traditionally used to score microsatellite genotypes, enable rapid evaluation and editing of automatically scored genotypes. megasat is written in Perl, runs on Windows, Mac OS X and Linux systems, and includes a simple graphical user interface. We demonstrate megasat using data from guppy, Poecilia reticulata. We genotype 1024 guppies at 43 microsatellites per run on an Illumina MiSeq sequencer. We evaluated the accuracy of automatically called genotypes using two methods, based on pedigree and repeat genotyping data, and obtained estimates of mean genotyping error rates of 0.021 and 0.012. In both estimates, three loci accounted for a disproportionate fraction of genotyping errors; conversely, 26 loci were scored with 0–1 detected error (error rate ≤0.007). Our results show that with appropriate selection of loci, automated genotyping of microsatellite loci can be achieved with very high throughput, low genotyping error and very low genotyping costs.     

Estimating and accounting for genotyping errors in RAD‐seq experiments

In non‐model organisms, evolutionary questions are frequently addressed using reduced representation sequencing techniques due to their low cost, ease of use, and because they do not require genomic resources such as a reference genome. However, evidence is accumulating that such techniques may be affected by specific biases, questioning the accuracy of obtained genotypes, and as a consequence, their usefulness in evolutionary studies. Here, we introduce three strategies to estimate genotyping error rates from such data: through the comparison to high quality genotypes obtained with a different technique, from individual replicates, or from a population sample when assuming Hardy‐Weinberg equilibrium. Applying these strategies to data obtained with Restriction site Associated DNA sequencing (RAD‐seq), arguably the most popular reduced representation sequencing technique, revealed per‐allele genotyping error rates that were much higher than sequencing error rates, particularly at heterozygous sites that were wrongly inferred as homozygous. As we exemplify through the inference of genome‐wide and local ancestry of well characterized hybrids of two Eurasian poplar (Populus) species, such high error rates may lead to wrong biological conclusions. By properly accounting for these error rates in downstream analyses, either by incorporating genotyping errors directly or by recalibrating genotype likelihoods, we were nevertheless able to use the RAD‐seq data to support biologically meaningful and robust inferences of ancestry among Populus hybrids. Based on these findings, we strongly recommend carefully assessing genotyping error rates in reduced representation sequencing experiments, and to properly account for these in downstream analyses, for instance using the tools presented here.     

How to track and assess genotyping errors in population genetics studies

Genotyping errors occur when the genotype determined after molecular analysis does not correspond to the real genotype of the individual under consideration. Virtually every genetic data set includes some erroneous genotypes, but genotyping errors remain a taboo subject in population genetics, even though they might greatly bias the final conclusions, especially for studies based on individual identification. Here, we consider four case studies representing a large variety of population genetics investigations differing in their sampling strategies (noninvasive or traditional), in the type of organism studied (plant or animal) and the molecular markers used [microsatellites or amplified fragment length polymorphisms (AFLPs)]. In these data sets, the estimated genotyping error rate ranges from 0.8% for microsatellite loci from bear tissues to 2.6% for AFLP loci from dwarf birch leaves. Main sources of errors were allelic dropouts for microsatellites and differences in peak intensities for AFLPs, but in both cases human factors were non-negligible error generators. Therefore, tracking genotyping errors and identifying their causes are necessary to clean up the data sets and validate the final results according to the precision required. In addition, we propose the outline of a protocol designed to limit and quantify genotyping errors at each step of the genotyping process. In particular, we recommend (i) several efficient precautions to prevent contaminations and technical artefacts; (ii) systematic use of blind samples and automation; (iii) experience and rigor for laboratory work and scoring; and (iv) systematic reporting of the error rate in population genetics studies.     

Consequences of genotyping errors for estimation of clonality: a case study on <Emphasis Type="Italic">Populus euphratica</Emphasis> Oliv. (Salicaceae)

A study including eight microsatellite loci for 1,014 trees from seven mapped stands of the partially clonal Populus euphratica was used to demonstrate how genotyping errors influence estimates of clonality. With a threshold of 0 (identical multilocus genotypes constitute one clone) we identified 602 genotypes. A threshold of 1 (compensating for an error in one allele) lowered this number to 563. Genotyping errors can seemingly merge (type 1 error), split really existing clones (type 2), or convert a unique genotype into another unique genotype (type 3). We used context information (sex and spatial position) to estimate the type 1 error. For thresholds of 0 and 1 the estimate was below 0.021, suggesting a high resolution for the marker system. The rate of genotyping errors was estimated by repeated genotyping for a cohort of 41 trees drawn at random (0.158), and a second cohort of 40 trees deviating in one allele from another tree (0.368). For the latter cohort, most of these deviations turned out to be errors, but 8 out of 602 obtained multilocus genotypes may represent somatic mutations, corresponding to a mutation rate of 0.013. A simulation of genotyping errors for populations with varying clonality and evenness showed the number of genotypes always to be overestimated for a system with high resolution, and this mistake increases with increasing clonality and evenness. Allowing a threshold of 1 compensates for most genotyping errors and leads to much more precise estimates of clonality compared with a threshold of 0. This lowers the resolution of the marker system, but comparison with context information can help to check if the resolution is sufficient to apply a higher threshold. We recommend simulation procedures to investigate the behavior of a marker system for different thresholds and error rates to obtain the best estimate of clonality.     

Incorporating Genotyping Error Into Non-Invasive DNA-Based Mark—Recapture Population Estimates

ABSTRACT Use of non-invasive sources of DNA, such as hair or scat, to obtain a genetic mark for population estimates is becoming commonplace. Unfortunately, with such marks, potentials for genotyping errors and for the shadow effect have resulted in use of many loci and amplification of each specimen many times at each locus, drastically increasing time and cost of obtaining a population estimate. We proposed a method, the Genotyping Uncertainty Added Variance Adjustment (GUAVA), which statistically adjusts for genotyping errors and the shadow effect, thereby allowing use of fewer loci and one amplification of each specimen per locus. Using allele frequencies and estimates of genotyping error rates, we determined, for each pair of specimens, the probability that the pair was obtained from the same individual, whether or not their observed genotypes match. Using these probabilities, we reconstructed possible capture history matrices and used this distribution to obtain a population estimate. With simulated data, we consistently found our estimates had lower bias and smaller variance than estimates based on single amplifications in which genotyping error was ignored and that were comparable to estimates based on data free of genotyping errors. We also demonstrated the method on a fecal DNA data set from a population of red wolves (Canis rufus). The GUAVA estimate based on only one amplification genotypes compares favorably to the estimate based on consensus genotypes. A program to conduct the analysis is available from the first author for UNIX or Windows platforms. Application of GUAVA may allow for increased accuracy in population estimates at reduced cost.     

Evaluation of genotyping concordance for commercial bovine SNP arrays using quality‐assurance samples

SNP arrays are widely used in genetic research and agricultural genomics applications, and the quality of SNP genotyping data is of paramount importance. In the present study, SNP genotyping concordance and discordance were evaluated for commercial bovine SNP arrays based on two types of quality assurance (QA) samples provided by Neogen GeneSeek. The genotyping discordance rates (GDRs) between chips were on average between 0.06% and 0.37% based on the QA type I data and between 0.05% and 0.15% based on the QA type II data. The average genotyping error rate (GER) pertaining to single SNP chips, based on the QA type II data, varied between 0.02% and 0.08% per SNP and between 0.01% and 0.06% per sample. These results indicate that genotyping concordance rate was high (i.e. from 99.63% to 99.99%). Nevertheless, mitochondrial and Y chromosome SNPs had considerably elevated GDRs and GERs compared to the SNPs on the 29 autosomes and X chromosome. The majority of genotyping errors resulted from single allotyping errors, which also included the opposite instances for allele ‘dropout’ (i.e. from AB to AA or BB). Simultaneous allotyping errors on both alleles (e.g. mistaking AA for BB or vice versa) were relatively rare. Finally, a list of SNPs with a GER greater than 1% is provided. Interpretation of association effects of these SNPs, for example in genome‐wide association studies, needs to be taken with caution. The genotyping concordance information needs to be considered in the optimal design of future bovine SNP arrays.     

Detection rates for genotyping errors in SNPs using the trio design

One well-known approach for the analysis of transmission-disequilibrium is the investigation of single nucleotide polymorphisms (SNPs) in trios consisting of an affected child and its parents. Results may be biased by erroneously given genotypes. Various reasons, among them sample swap or wrong pedigree structure, represent a possible source for biased results. As these can be partly ruled out by good study conditions together with checks for correct pedigree structure by a series of independent markers, the remaining main cause for errors is genotyping errors. Some of the errors can be detected by Mendelian checks whilst others are compatible with the pedigree structure. The extent of genotyping errors can be estimated by investigating the rate of detected genotyping errors by Mendelian checks. In many studies only one SNP of a specific genomic region is investigated by TDT which leaves Mendelian checks as the only tool to control genotyping errors. From the rate of detected errors the true error rate can be estimated. Gordon et al. [Hum Hered 1999;49:65-70] considered the case of genotyping errors that occur randomly and independently with some fixed probability for the wrong ascertainment of an allele. In practice, instead of single alleles, SNP genotypes are determined. Therefore, we study the proportion of detected errors (detection rate) based on genotypes. In contrast to Gordon et al., who reported detection rates between 25 and 30%, we obtain higher detection rates ranging from 39 up to 61% considering likely error structures in the data. We conclude that detection rates are probably substantially higher than those reported by Gordon et al.     

Genetic mapping in the presence of genotyping errors

Genetic maps are built using the genotypes of many related individuals. Genotyping errors in these data sets can distort genetic maps, especially by inflating the distances. We have extended the traditional likelihood model used for genetic mapping to include the possibility of genotyping errors. Each individual marker is assigned an error rate, which is inferred from the data, just as the genetic distances are. We have developed a software package, called TMAP, which uses this model to find maximum-likelihood maps for phase-known pedigrees. We have tested our methods using a data set in Vitis and on simulated data and confirmed that our method dramatically reduces the inflationary effect caused by increasing the number of markers and leads to more accurate orders.