The protocol presents a series of best practice protocols for the collection of bone powder from eight recommended anatomical sampling locations (specific locations on a given skeletal element) across five different skeletal elements from medieval individuals (radiocarbon dated to a period of ca. 1040-1400 CE, calibrated 2-sigma range).
The methods presented here seek to maximize the chances for the recovery of human DNA from ancient archaeological remains while limiting input sample material. This was done by targeting anatomical sampling locations previously determined to yield the highest amounts of ancient DNA (aDNA) in a comparative analysis of DNA recovery across the skeleton. Prior research has suggested that these protocols maximize the chances for the successful recovery of ancient human and pathogen DNA from archaeological remains. DNA yields were previously assessed by Parker et al. 2020 in a broad survey of aDNA preservation across multiple skeletal elements from 11 individuals recovered from the medieval (radiocarbon dated to a period of circa (ca.) 1040-1400 CE, calibrated 2-sigma range) graveyard at Krakauer Berg, an abandoned medieval settlement near Peißen Germany. These eight sampling spots, which span five skeletal elements (pars petrosa, permanent molars, thoracic vertebra, distal phalanx, and talus) successfully yielded high-quality ancient human DNA, where yields were significantly greater than the overall average across all elements and individuals. Yields were adequate for use in most common downstream population genetic analyses. Our results support the preferential use of these anatomical sampling locations for most studies involving the analyses of ancient human DNA from archaeological remains. Implementation of these methods will help to minimize the destruction of precious archaeological specimens.
The sampling of ancient human remains for the purposes of DNA recovery and analysis is inherently destructive1,2,3,4. The samples themselves are precious specimens and morphological preservation should be preserved wherever possible. As such, it is imperative that sampling practices be optimized to both avoid unnecessary destruction of irreplaceable material and to maximize the probability of success. Current best practice techniques are based on a small cohort of studies limited to either forensic surveys5,6, studies of ancient specimens where the development of optimal sampling is not the direct aim of the study7, or dedicated studies utilizing either non-human remains8 or targeting a very small selection of anatomical sampling locations (used here to denote a specific area of a skeletal element from which bone powder, for use in downstream DNA analyses, was generated)9,10. The sampling protocols presented here were optimized in the first large-scale systematic study of DNA preservation across multiple skeletal elements from multiple individuals11. All samples stemmed from skeletal elements recovered from 11 individuals excavated from the church graveyard of the abandoned medieval settlement of Krakauer Berg near Peißen, Saxony-Anhalt, Germany (see Table 1 for detailed sample demographics) and, as such, may need modification for use with samples outside of this geographical/temporal range.
Individual | Sex | Estimated age at death | 14C dates (CE, Cal 2-sigma) |
KRA001 | Male | 25-35 | 1058-1219 |
KRA002 | Female | 20-22 | 1227-1283 |
KRA003 | Male | 25 | 1059-1223 |
KRA004 | Male | 15 | 1284-1392 |
KRA005 | Male | 10-12 | 1170-1258 |
KRA006 | Female | 30-40 | 1218-1266 |
KRA007 | Female | 25-30 | 1167-1251 |
KRA008 | Male | 20 | 1301-1402 |
KRA009 | Male | Unknown | 1158-1254 |
KRA010 | Male | 25 | 1276-1383 |
KRA011 | Female | 30-45 | 1040-1159 |
Table 1: Genetically determined sex, archaeologically determined estimated age at death, and radiocarbon dating (14C Cal 2-sigma) for all the 11 individuals sampled. This table has been adapted from Parker, C. et al. 202011.
These protocols allow for a relatively straightforward and efficient generation of bone powder from eight anatomical sampling locations across five skeletal elements (including the pars petrosa) with limited laboratory-induced DNA contamination. Of these five skeletal elements, seven anatomical sampling locations found on four skeletal elements have been determined to be viable alternatives to the destructive sampling of the petrous pyramid11,12. These include the cementum, dentin, and pulp chamber of permanent molars; cortical bone gathered from the superior vertebral notch as well as from the body of thoracic vertebrae; cortical bone stemming from the inferior surface of the apical tuft and shaft of the distal phalanges; and the dense cortical bone along the exterior portion of the tali. While there are several widely applied methods for the sampling of the pars petrosa4,12,13,14, dentin, and the dental pulp chamber1,2,15, published methods describing the successful generation of bone powder from the cementum16, vertebral body, inferior vertebral notch, and talus can be difficult to obtain. As such, here we demonstrate optimized sampling protocols for the petrous pyramid (step 3.1); cementum (step 3.2.1), dentin (step 3.2.2), and dental pulp (step 3.2.3) of adult molars; cortical bone of the vertebral body (step 3.3.1) and superior vertebral arch (step 3.3.2); the distal phalanx (step 3.4); and the talus (step 3.5) in order to make the effective use of these skeletal elements for both aDNA and forensic research more widely accessible.
All research presented herein was performed in compliance with the guidelines set forth by the Max Planck Institute for the Science of Human History, Jena, Germany for working with ancient human remains. Before performing any steps of this protocol ensure to adhere to all local/state/federal ethical requirements pertaining to both obtaining permission for the scientific study and use of human remains for destructive sampling in your area. All procedures/chemical storage should be performed according to individual institutional safety guidelines.
1. Considerations before sample processing
2. Pretreatment
3. Bone powder generation
NOTE: The following protocols are intended for use in DNA extraction following the Dabney et al. 2019 protocol26.
Figure 1: Temporal bone including the pars petrosa. (A) Sample pre-cutting showing the locations of the petrous pyramid and the sulcus petrosa. (B) Petrous portion post-cutting highlighting the dense areas to be drilled. Please click here to view a larger version of this figure.
Figure 2: Permanent molar pre-sampling. (A) Pre-treated molar prior to sampling, showing crown, cementum (yellowish layer of the root), and the cutting site at the cemento-enamel junction. (B) The same molar post-cementum collection, showing the cut site at the cemento-enamel junction. (C) Molar post-cutting and sampling showing anatomical sampling locations for the dental pulp chamber and dentin within the crown. Please click here to view a larger version of this figure.
Figure 3: Vertebral body and superior vertebral arch cortical bone anatomical sampling locations of the thoracic vertebra. Please click here to view a larger version of this figure.
Figure 4: Distal phalanx showing the locations of dense cortical bone along the shaft and inferior side of the apical tuft. Please click here to view a larger version of this figure.
Figure 5: Sampling area of the talus for cortical bone recovery. Please click here to view a larger version of this figure.
NOTE: The talus has very little cortical bone (a thin outer layer). The material should not only be collected from the surface but also the underlying dense layer of cancellous bone.
In a separate study11, DNA was extracted from bone powder generated from each anatomical sampling location in 11 individuals, using a standard DNA extraction protocol optimized for short fragments from calcified tissue2. Single-stranded libraries were then produced28 and sequenced on a HiSeq 4000 (75 bp paired-end) to a depth of ~20,000,000 reads per sample. The resulting sequence data was then evaluated for endogenous human DNA content using the EAGER pipeline29 (BWA settings: Seed length of 32, 0.1 mismatch penalty, mapping quality filter of 37). All representative results are reported using the same metrics as Parker et al. 202011 for consistency. Libraries from the powdered portions of the pars petrosa yielded, on average, higher endogenous DNA than any of the other 23 anatomical sampling locations surveyed (Figure 6A–B). The seven additional anatomical sampling locations presented in this protocol (the cementum, first pass of the dental pulp chamber, and dentin of permanent molars; cortical bone from the vertebral body and superior vertebral arch of the thoracic vertebra; cortical bone from the apical tuft of the distal phalanx; and cortical bone from the neck of the talus) produced the next highest yields (with no statistical significance between these anatomical sampling locations; Figure 6A–B; Supplemental File 1: EndogenousDNAPreCap). These alternative locations all consistently produced DNA yields adequate for standard population genetics analyses such as mitochondrial analyses and single nucleotide polymorphism (SNP) analyses. Duplication rates in libraries stemming from all anatomical sampling locations were low (cluster factors < 1.2 on average, calculated as the ratio of all mapping reads to unique mapping reads, Table 2; Supplemental File 1: ClusterFactor), indicating that all libraries screened were of very high complexity. Similarly, average exogenous human DNA contamination estimates were low, averaging < 2% (X chromosome contamination in males, n = 7, as reported by the ANGSD30 pipeline) in all anatomical sampling locations except for the superior vertebral arch (average estimated contamination: 2.11%, with one sample removed as an outlier; KRA005: 19.52%, see Table 2; Supplemental File 1: Xcontamination). Average fragment length (after filtering to remove all reads < 30 bp) was lowest in the material collected from the dental pulp chamber and dentin, with no significant variation among other anatomical sampling locations (55.14 bp and 60.22 bp, respectively in comparison to an average median of 62.87, pair-wise p-values < 0.019, Table 2; Supplemental File 1: AvgFragLength). Additionally, the teeth and thoracic vertebrae each contain multiple anatomical sampling locations where high endogenous DNA recovery was observed, making them particularly suitable as alternatives to the pars petrosa.
Figure 6: Human DNA content for all screened samples. Black lines represent the overall mean, while red lines represent the median (solid: human DNA proportion, dashed: mapped human reads per million reads generated). Individual anatomical sampling locations with an average human DNA proportion higher than the overall mean (8.16%) are colorized in all analyses. (A) The proportion of reads mapping to the hg19 reference genome. The blue dashed line represents the theoretical maximum given the pipeline's mapping parameters (generated using Gargammel31 to simulate a random distribution of 5,000,000 reads from the hg19 reference genome with simulated damage). Individual means (black X) and medians (red circle) are reported for those samples with a higher average human DNA proportion than the overall mean. Confidence intervals indicate upper and lower bounds excluding statistical outliers. (B) The number of unique reads mapping to the hg19 reference genome per million reads of sequencing effort (75 bp paired end). Confidence intervals indicate upper and lower bounds excluding statistical outliers. This figure has been adapted from Parker, C. et al. 202011. Please click here to view a larger version of this figure.
Table 2: Average duplication levels (mapping reads/unique reads), average and median fragment lengths, and X chromosome contamination estimates for all anatomical sampling locations. Error reported as the standard error of the mean. This table has been adapted from Parker, C. et al. 202011.
Sampling location | Average duplication factor (# mapped reads /# unique mapped reads) | Average fragment length in bp | Average estimated proportion of X chromosome contamination |
Petrous pyramid | 1.188 ± 0.006 | 65.40 ± 1.36 | 0.000 ± 0.003 |
Cementum | 1.197 ± 0.028 | 67.28 ± 1.76 | 0.011 ± 0.003 |
Dentin | 1.188 ± 0.061 | 60.22 ± 2.37 | 0.002 ± 0.007 |
Pulp | 1.179 ± 0.024 | 55.14 ± 2.90 | 0.013 ± 0.006 |
Distal phalanx | 1.191 ± 0.049 | 65.95 ± 1.08 | 0.013 ± 0.005 |
Vertebral body | 1.194 ± 0.037 | 66.14 ± 1.03 | 0.008 ± 0.003 |
Superior vertebral arch | 1.19 ± 0.017 | 63.02 ± 1.23 | 0.021 ± 0.009* |
Talus | 1.198 ± 0.010 | 68.20 ± 1.24 | 0.011 ± 0.003 |
*Sample KRA005 removed as an outlier at 0.1952 |
Code availability
All analyses programs and R modules used in the analyses of this manuscript are freely available from their respective authors. All custom R code is available by request.
Data availability
All raw data used in the calculation of representative results is freely available in the European Nucleotide Archive ENA data repository (accession number PRJ-EB36983) or supplemental materials of Parker, C. et al.11.
Supplemental File 1. Please click here to download this File.
Current practice in ancient human population genetics is to preferentially sample from the pars petrosa (step 2.1) whenever possible. However, the pars petrosa can be a difficult sample to obtain, as it is highly valued for a myriad of skeletal assessments (e.g., population history32, the estimation of fetal age at death33, and sex determination34), and, historically, sampling of the pars petrosa for DNA analysis can be highly destructive3,4 (including the protocol presented here, although new, minimally invasive protocols13,14 have now been widely adopted to alleviate this concern). This is compounded by the fact that, until very recently, a large-scale, systematic study of human DNA recovery across the skeleton had not been attempted11, making finding an appropriate sampling strategy when the petrous pyramid is unavailable challenging.
The protocols presented here help to alleviate that challenge by providing a set of optimized procedures for DNA sampling from archaeological/forensic skeletal remains including the pars petrosa as well as seven alternate anatomical sampling locations across four additional skeletal elements. The critical steps included are all intended to minimize the possibility of DNA loss/damage due to either inefficient sampling (steps 2.1.6 and 3.2.1.3) or overheating of samples during drilling/cutting (step 3.1.6). Additionally, it has been noted throughout the protocol that it may be necessary to modify/omit the pre-treatment steps to ensure the best performance in highly degraded samples. It should also be noted that even among the selected elements presented here, there remain several possible alternative sampling techniques (particularly for the pars petrosa13,14), as well as ample room for further optimization of the underexploited anatomical sampling locations presented here (i.e., the talus: step 2.5 and the vertebrae: step 2.3).
It is also important to keep in mind that these protocols have been designed and tested using ancient juvenile-adult remains of high quality (good morphological preservation) for the purposes of endogenous human DNA analyses. The results presented may not extend to more highly degraded materials, other preservation contexts, infant remains, non-human remains, or studies of pathogens or the microbiome, as a greater exploration into the use of these protocols in additional contexts is still needed. Additionally, the alternative skeletal elements presented here (the teeth, vertebrae, distal phalanx, and tali) may be challenging to assign to a single individual among commingled remains, necessitating sampling from multiple elements to ensure a single origin. Despite these limitations, making these protocols widely available can help alleviate some of the heterogeneity surrounding sample selection and processing by providing a generalized and quantitatively optimized framework for use in a wide range of future aDNA/forensic studies on human remains.
The authors have nothing to disclose.
The authors would like to thank the laboratory staff of the Max Planck Institute for the Science of Human History for their help in the development and implementation of these protocols. This work would not have been possible without the input and hard work of Dr. Guido Brandt, Dr. Elizabeth Nelson, Antje Wissegot, and Franziska Aron. This study was funded by the Max Planck Society, the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program under grant agreements No 771234 – PALEoRIDER (WH, ABR) and Starting Grant No. 805268 CoDisEASe (to KIB).
#16 Dental Drill Bit | NTI | H1-016-HP | example drilling bit |
0.6 mm scroll saw blade | Fisher Scientific | 50-949-097 | blade for Jewellers Saw |
22mm diamond cutting wheel | Kahla | SKU 806 104 358 514 220 | Dremel cutting attachment |
Commercial Bleach | Fisher Scientific | NC1818018 | |
Control Company Ultra-Clean Supreme Aluminum Foil | Fisher Scientific | 15-078-29X | |
DNA LoBind Tubes (2 mL) | Eppendorf | 22431048 | |
Dremel 225-01 Flex Shaft Attachment | Dremel | 225-01 | Dremel flexible extension |
Dremel 4300 Rotary Tool | Dremel | 4300 | Example drill |
Dremel collet and nut kit | Dremel | 4485 | Adapters for various Dremel tool attachments/bits |
Eagle 33 Gallon Red Biohazard Waste Bag | Fisher Scientific | 17-988-501 | |
Eppendorf DNA LoBind 2 mL microcentrifuge tube | Fisher Scientific | 13-698-792 | |
Ethanol (Molecular Biology Grade) | Millipore Sigma | 1.08543 | |
FDA approved level 2 Surgical Mask | Fisher Scientific | 50-206-0397 | PPE |
Fisherbrand Comfort Nitrile Gloves | Fisher Scientific | 19-041-171X | PPE |
Fisherbrand Safety Glasses | Fisher Scientific | 19-130-208X | PPE |
Granger Stationary Vise | Fisher Scientific | NC1336173 | benchtop vise |
Invitrogen UltraPure DNase/Rnase free distilled water | Fisher Scientific | 10-977-023 | |
Jewellers Saw | Fisher Scientific | 50-949-231 | |
Kimwipes | Sigma-Aldritch | Z188956 | |
Labconco Purifier Logic Biosafety cabinet | Fisher Scientific | 30-368-1101 | |
LookOut DNA Erase | Millipore Sigma | L9042-1L | |
Medium weighing boat | Heathrow Scientific | HS120223 | |
MSC 10pc plier/clamp set | Fisher Scientific | 50-129-5352 | Miscellaneous clamps/vise grips for securely holding samples while drilling/cutting |
Sartorius Quintix Semi-Micro Balance | Fisher Scientific | 14-560-019 | enclosed balance |
Tyvek coveralls with hood | Fisher Scientific | 01-361-7X | PPE |
Weigh paper | Heathrow Scientific | HS120116 |