Collagen of ancient bones gives an indication of endogenous DNA preservation based on next-generation sequencing technology

Abstract

Ancient genome analysis has become an indispensable tool in studies of human population history and evolution since the breakthrough of whole-genome sequencing technology. The problem remains, however, that ancient genomes cannot be analyzed without crushing non-small pieces of precious specimens; moreover, in many cases, there is insufficient DNA remaining in the pieces of sample to obtain whole-genome sequences. In previous studies, therefore, a couple of indicators (e.g. racemization ratios) have been proposed to estimate the endogenous DNA in ancient samples. However, these studies have used polymerase chain reaction (PCR) to test whether endogenous DNA remains, but this has proved inadequate because the success or failure of PCR amplification does not necessarily reflect the DNA remaining. To assess the amount of endogenous DNA, we use the ratios of reads generated by next-generation sequencing (NGS) mapped to the human reference genome sequence. We investigated 40 human remains excavated from three shell-mound sites of the late to final Jomon culture. The associations between the environmental/molecular factors and the mapping ratios (MRs) were examined. There were no significant associations between the environmental factors and MRs, or between the collagen residual ratios (CRRs) and the MRs. However, we found a significant association between CRRs in rib bones and MRs. The weight of bone required to measure residual collagen is much less than that required to obtain the DNA necessary for NGS analysis, and the process of measuring CRRs is always involved in dating. Hence, we propose the collagen in the ribs as a good indicator for successful ancient genome analyses.

Introduction

DNA extracted from ancient biological remains provides important information on the evolution and diversity of organisms in the past. Initially, analyses of ancient DNA sequences relied on the dideoxy (Sanger) method through polymerase chain reaction (PCR) amplification (Pääbo et al., 1988, 1989; Hagelberg et al., 1989; Lawlor et al., 1991; Kurosaki et al., 1993; Oota et al., 1995; Shinoda and Kanai, 1999). PCR-based Sanger sequencing, however, provides only a limited amount of genome information. Whole-genome sequencing became quicker and cheaper with the advent of next-generation sequencing (NGS) in the 21st century, and this technology has been used to analyze ancient genomes of archaic hominins (e.g. Green et al., 2006, 2010; Krause et al., 2010; Reich et al., 2010; Meyer et al., 2012; Prüfer et al., 2014) and modern human genomes (e.g. Fu et al., 2014; Rasmussen et al., 2010; Kanzawa-Kiriyama et al., 2017, 2019; McColl et al., 2018; Gakuhari et al., 2020). This research has provided us with a wealth of information about archaic hominins and the peopling history of modern humans spreading from Africa to other continents.

Despite technological advances in DNA sequencing, genome analyses of paleobiological subjects are fraught with difficulties: ancient DNA undergoes chemical modifications postmortem in organisms, fragmenting and reducing the number of molecules. Soft and hard tissues of organisms are degraded by bacteria and fungi that enter through the surrounding soil in which the remains are buried, and some environmental factors, such as pH, temperature, exposure to water (Lindahl, 1993; Hedges and Millard, 1995), and humidity (Pinhasi et al., 2015), affect the rate of DNA molecular degradation. Because the DNA strand is fragmented and the number of DNA molecules is reduced dramatically through the unavoidable processes mentioned above, ancient samples to be analyzed do not always contain sufficient amounts of DNA for genome sequencing. This difficulty leads to another difficulty in selecting samples for ancient genome analysis. Even if valuable ancient samples are crushed for DNA extraction, they could be wasted if the DNA is not available.

For these reasons, a non-crushing or a nearly non-crushing indicator for estimating DNA preservation has been sought in previous studies to minimize damaging valuable biological remains to be analyzed. Initially, the rate of racemization of amino acids was used to assess the state of DNA preservation in ancient biological remains (Poinar et al., 1996). Subsequently, the collagen residual ratios (CRRs) and crystallinity of hydroxyapatite in the specimens were reported as suitable indicators (Götherström et al., 2002). Both the racemization rate and the collagen residual rate gave certain criteria for DNA preservation, but the presence/absence of DNA was determined on the basis of successful or unsuccessful PCR amplification. However, the possibility of unsuccessful amplification by PCR cannot be excluded. For example, if the primer cannot anneal due to mutations or the binding of contaminants at the primer binding site, or if the DNA fragments are shorter than the target sequence, PCR amplification will not proceed effectively. There is also the possibility of false-positive amplification caused by contamination. Hence, a method that avoids the above problems in assessing the state of DNA preservation in ancient specimens is required.

Here we show an NGS-based method to obtain a more quantitative indication of DNA preservation exclusively in human skeletal remains. We consider the rate at which reads from shotgun sequencing map to the human reference sequence as the percentage of DNA remaining in the ancient sample. We first compared whether environmental factors (in/out of a shell stratum and calibrated radiocarbon dates) or the molecular factor (CRR) correlated with the mapping ratio (MR). The results showed that CRRs, as expected from previous studies, correlated best with MRs, although no association was shown in both environmental factors. The association was statistically significant, especially in rib bones. This could be useful to assess the amount of DNA remaining and suggests a way to develop a method that avoids unnecessary crushing of precious ancient human bones.

Materials and Methods

Archaeological samples

We examined 40 individuals excavated from three shell-mound sites (15 from Gionbara [GB], 7 from Saihiro [SH], and 18 from Kikumatenaga [KT]) located in Ichihara City, Chiba Prefecture, and within 10 km of each other (Figure 1). Based on the chronological age of earthenware (Habu, 2004), the GB and SH sites were assigned to the late to final Jomon period (4000–2300 BP), whereas the KT site was assigned to the late Jomon period (4000–3000 BP).

Figure 1.

The geographical location of the three archaeological sites: Gionbara, Kikumatenaga, and Saihiro shell mounds.

All human skeletal remains from the SH and KT sites were excavated from the shell stratum. At the GB site, four individuals were excavated from the shell stratum; ten were excavated from the non-shell stratum; one was unknown (Kondo, 1987; Oshizawa, 1999; Sakurai, 2005). Whether a stratum is shell or non-shell affects the pH of the surrounding sediments.

Gelatin extraction and ¹⁴C dating

Each sample was subjected to collagen extraction from one different part (22 ribs, 11 skulls; 5 limbs; 1 ilium). Gelatin extraction was based on protocols used in previous studies (Longin, 1971; Yoneda et al., 2002). Bone samples were cleaned by sandblasting and ultrasonic cleaning for 10 min. They were then demineralized for 40 h with 0.4 M HCl and neutralized for 23 h 50 min with pure water. After the first neutralization, samples were treated with 0.1 M NaOH for 1 h and neutralized for 3 h 30 min. After the second neutralization, samples were gelatinized for 41 h at 90°C with 0.0001 M HCl (pH 4.0). Suction filtration was performed using Whatman GF/F filters. Filtrated samples were freeze-dried, and the amount of gelatin was weighed; this weight was considered to be the collagen content. Radiocarbon dating was measured by accelerator mass spectrometry at The University Museum, The University of Tokyo. Dates were referenced to IntCal20 and Marine 20 (Reimer et al., 2020; Heaton et al., 2020), and calibrated ages were estimated using OxCal 4.4 software (Bronk Ramsey, 2009). The marine reservoir effect, in which the apparent radiocarbon dates become older due to seafood exploitation, was corrected by evaluating the effect of marine carbon based on δ¹³C. Using the mean values (–22.6‰ and –10.9‰) of land mammal and marine fish bones excavated from shell middens in Chiba Prefecture as end-members, we estimated the marine contribution and mixed IntCal20 and Marine20, assuming an error of 5%. A regional correction value (ΔR) for Tokyo Bay was adapted to Marine20 (–9837 years) based on a shell collected in 1882 AD (534 ± 36 BP; Yoshida et al., 2010). The median of the probability distribution of dating was used as the estimated age.

DNA extraction

Petrous bones from 40 individuals were used in this study. DNA was extracted following the protocols described in previous studies (Gamba et al., 2014; Gakuhari et al., 2020). The petrous bone samples were cut with sterile disc cutters and drills to obtain pieces weighing about 100 mg. The bone pieces were placed in in 5 ml tubes of DNA Lobind and predigested for 15 min at 900 r.p.m. in a Thermomixer (Eppendorf, Hamburg, Germany) with lysis buffer (2 ml) containing 20 mM Tris–HCl (pH 7.4), 0.7% N-lauroylsarcosine sodium salt solution, 0.5 M EDTA (pH 8.0), and 1.2 U/ml recombinant Proteinase K. After the supernatant was transferred, the lysis buffer was added to the sample, which was then placed in a Thermomixer at 900 r.p.m. and 60°C overnight (>16 h). After digestion, the tube was centrifuged for 10 min at 3500 r.p.m. The supernatant was diluted with 10 mM TE buffer added to the filter unit (Amicon Ultra 4, 10 kDa), and centrifuged for 30 min at 3500 r.p.m., or until a final volume of 100 μl was reached. The final volume was added to a Qiagen Mini elute column in a 2 ml tube and purified according to the manufacturer’s instructions, apart from adding Tween 20 to EB buffer preheated at 60°C.

All experiments were performed in a clean room used exclusively for ancient genome experiments in the Department of Biological Sciences, Graduate School of Science, The University of Tokyo. Before and after use, the room was illuminated with UV light. The clean room in the working state was kept at a positive pressure and air without modern DNA was supplied through a HEPA filter. The researchers always wore a body suit, face mask, and gloves. The tools and bone surfaces were exposed to UV light.

Library construction

The 40 libraries were constructed with a NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA) according to the manufacturer’s instructions, apart from changing the adaptor dilution from 25-fold to 10-fold, and the first bead addition from 40 μl to 90 μl; similarly, the second bead addition was changed from 20 μl to 90 μl. The concentration of DNA extracts and libraries was measured with a Qubit (ThermoFisher, Waltham, MA, USA). The fragment length of DNA extracts and libraries was checked with a Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). The sequencing was carried out in MiSeq (Illumina, San Diego, CA, USA). The libraries were sequenced on a flow cell using an MiSeq Reagent Kit v. 3 150-cycle.

Raw sequencing data processing

The FASTQ files generated by MiSeq were processed by the following pipeline. AdaptorRemoval v. 2.2.2 (Schubert et al., 2016) was used to trim ambiguous bases at termini (-trimns), low-quality bases at termini (-trimqualities), and short leads (-minlength 35), and to combine paired-end reads into a consensus sequence (-collapse). BWA v. 0.7.16 (Li and Durbin, 2009) was used to map combined reads to human genome reference hg19. The MR (the ratio of the number of mapped reads divided by the number of reads post-AdaptorRemoval) was calculated by flagstat in SAMtools v. 1.10 (Danecek et al., 2021). CleanSam in Picard Tools v. 2.21.8 (https://broadinstitute.github.io/picard/) was used to soft-clip beyond-end-of reference alignments. Duplicate reads were removed by MarkDuplicates in Picard Tools with the parameter REMOVE_DUPLICATES set to true. The substitution pattern of DNA molecules in libraries was checked by mapDamage2 v. 2.2.0 (Jónsson et al., 2013).

Statistical analysis

The Mann–Whitney U-test was used to compare MRs between bones in the shell strata and those not in the shell strata. Linear regression analysis was conducted to examine the association of ¹⁴C dates with MRs, and the association of CRRs with MRs. Furthermore, the association between CRRs and MRs was examined separately for the skulls and the ribs. The CRR was defined as gelatin weight divided by bone weight. Analysis of variance (ANOVA) was used to assess differences of the CRRs and the MRs between sites. The association of CRRs with MRs was also assessed by a linear model with two dummy variables representing three sites (SH, KT, or GB). In the analysis, GB was used as the reference. Generalized variance inflation factors (GVIFs) were used to check the multicollinearity for CRRs and sites. The relationship between CRRs of ribs and MRs for each site was also assessed by a linear model. The R package “coin” was used to carry out the Mann–Whitney U-test. The R packages “tidyverse” and “forcats” were used to carry out the regression analysis. The R package “car” was used to carry out ANOVA and to calculate the GVIFs.

Results

Table 1 summarizes the MRs, CRRs, and carbon-to-nitrogen (CN) ratios of all the samples. The MR is a proxy for the proportion of endogenous DNA based on the assumption that there is no contamination of modern DNA. For instance, we obtained 2.55 ng/μl of DNA and an MR of 0.01% from GB2-63, indicating that almost all DNA extracted from this sample was from non-human organisms. Among 40 samples, the best MR, 62.97%, was obtained from KT69. As observed in previous studies on ancient DNA, the MRs varied from each sample, with ranges of 0.01–44.52%, 9.03–48.64%, and 0.01–62.97% in GB, SH, and KT, respectively, showing typical patterns of deamination (Supplementary Figure 1). Radiocarbon dating was measured for 39 samples (1 out of 40 failed to be measured). The CRRs were 0.5–9.6% and the CN ratios were 2.9–3.6, indicating that the collagen was not denatured (DeNiro, 1985), except for 3 out of 39 samples. Because there was no relationship between the MR and the CN ratio, we used all the CRRs (including the three samples) in the subsequent analyses. We successfully measured radiocarbon dating for 35 samples but failed for 4 out of 39 samples because their gelatin was not extracted. The estimated dating was 5076–3424 calBP. The numbers of samples that were used for each test of an association between environmental and/or molecular factors and MRs are summarized in Table 2.

Table 1.

Summary of the measurments In DNA and collagen

Sample ID	Shell-mound site	DNA conc. (ng/μl)	Number of raw reads	Number of mapped reads	MR (%)	CRR (%)	CN	Bone parts for collagen analysis	Date (BP) estimated	Date (calBP) median	AMS-ID	Shell stratum
GB1-9	GB	3.16	2758827	67167	2.43	4.4	3.5	Limb	3548 ± 21	3843	TKA-22956	Out
GB3-15b	GB	3.08	513967	4051	0.79	1.6	3.6	Rib	3420 ± 23	3529	TKA-22989	NA
GB2-58	GB	2.20	541217	1499	0.28	2.5	3.3	Rib	3984 ± 24	4304	TKA-22954	Out
GB2-63	GB	2.55	1426936	133	0.01	0.5	4.3	Skull	NA	NA	NA	Out
GB3-3	GB	1.04	1499935	667812	44.52	7.8	3.4	Rib	3573 ± 21	3664	TKA-22962	In
GB3-8	GB	1.97	593504	11954	2.01	3.4	3.5	Rib	3779 ± 23	3933	TKA-22975	In
GB3-9	GB	0.53	1382453	257919	18.66	2.3	3.3	Rib	3738 ± 23	3951	TKA-22990	Out
GB3-10a	GB	0.59	1704665	122737	7.20	2.1	3.3	Limb	3764 ± 23	3924	TKA-23031	Out
GB3-11a	GB	0.60	1754314	27234	1.55	2.4	3.3	Rib	3643 ± 22	3752	TKA-22982	In
GB3-13a	GB	0.80	1734033	358	0.02	6.3	3.5	Skull	3907 ± 21	4133	TKA-22949	In
GB3-31	GB	2.32	3239859	308545	9.52	4.5	3.3	Rib	3935 ± 26	4182	TKA-22992	Out
GB3-32	GB	3.34	497359	53	0.01	NA	NA	NA	NA	NA	NA	Out
GB4-310e	GB	2.57	598741	71	0.01	1.2	3.5	Skull	3824 ± 27	4030	TKA-22967	Out
GB4-S006b	GB	2.14	692339	40254	5.81	4.2	3.3	Skull	3651 ± 22	3869	TKA-22978	Out
GB4-S022	GB	0.83	2232872	680235	30.46	1.3	3.4	Rib	3864 ± 24	4169	TKA-20259	Out
SH7-2	SH	1.63	613477	189192	30.84	5.0	3.3	Skull	3851 ± 27	4141	TKA-20252	NA
SH7-3	SH	2.16	648735	127469	19.65	1.1	3.5	Skull	3837 ± 24	4102	TKA-20253	NA
SH7-4	SH	0.81	663608	99138	14.94	6.0	3.4	Rib	3838 ± 24	4036	TKA-20254	NA
SH7-6	SH	1.42	660187	239282	36.24	5.1	3.3	Skull	3747 ± 25	3912	TKA-20255	NA
SH7-7	SH	0.99	1646380	800824	48.64	5.6	3.4	Rib	3999 ± 25	4279	TKA-20256	NA
SH7-10	SH	1.83	540642	48837	9.03	0.7	3.4	Skull	3302 ± 25	3424	TKA-20257	NA
SH7-12	SH	1.37	1887436	793518	42.04	4.9	3.4	Rib	3910 ± 26	4127	TKA-20258	NA
KT1-1	KT	0.33	11836	4731	39.97	2.6	3.7	Skull	4549 ± 22	5076	TKA-25369	NA
KT1-2	KT	0.47	1954232	232573	11.90	7.8	3.2	Skull	3738 ± 22	3905	TKA-25370	NA
KT2	KT	0.11	1999626	723778	36.20	4.0	3.2	Rib	3444 ± 21	3589	TKA-25368	NA
KT11-1	KT	0.34	2606777	363131	13.93	3.3	3.3	Limb	NA	NA	NA	NA
KT13-1	KT	0.60	7178	1	0.01	5.7	3.4	Limb	4093 ± 22	4429	TKA-25314	NA
KT14-1	KT	0.15	777858	186157	23.93	1.3	4.6	Skull	NA	NA	NA	NA
KT21	KT	0.06	958	84	8.77	9.6	3.3	Limb	NA	NA	NA	NA
KT52-2	KT	0.10	1798441	1045442	58.13	4.7	3.3	ilium	3897 ± 21	4076	TKA-24758	NA
KT53	KT	0.13	309850	3128	1.01	4.3	3.2	Rib	3706 ± 21	3793	TKA-25378	NA
KT57	KT	0.04	3074	1	0.03	4.3	3.2	Rib	3723 ± 21	3887	TKA-25377	NA
KT58	KT	0.89	39198	22678	57.85	6.6	3.2	Rib	3671 ± 22	3792	TKA-25387	NA
KT62	KT	0.54	3773	36	0.95	5.0	3.2	Rib	3760 ± 21	3944	TKA-25380	NA
KT63	KT	0.73	9930	40	0.40	3.1	3.2	Rib	3637 ± 21	3757	TKA-25381	NA
KT64	KT	0.20	13479	5628	41.75	6.9	3.2	Rib	3700 ± 22	3783	TKA-25375	NA
KT65	KT	0.45	1738998	34134	1.96	4.7	3.2	Rib	3708 ± 21	3886	TKA-25376	NA
KT68	KT	0.13	1526001	857922	56.22	5.3	3.2	Rib	3723 ± 22	3887	TKA-25373	NA
KT69	KT	0.10	33612	21165	62.97	4.4	3.2	Rib	3615 ± 22	3736	TKA-25374	NA
KT80	KT	0.18	54703	9083	16.60	6.3	3.2	Rib	3846 ± 21	3976	TKA-24744	NA

NA, data not available.

Table 2.

The numbers of samples for testing correlation with the factors

	Shell stratum	Dating	Collagen	Total
Gionbara (GB)	14	13	14	15
Saihiro (SH)	NA	7	7	7
Kikumatenaga (KT)	NA	15	18	18
Total	14	35	39	40

First, we examined an association between environmental factors and MRs. The previous studies reported that the bones were in relatively good condition when the remains were buried in a stratum that was slightly alkaline (Gordon and Buikstra, 1981; Stephan, 1997). In general, the sedimental pH of an in-shell-stratum is higher than that of an out-shell-stratum. We expected, therefore, that DNA extracted from bones in the shell stratum could be in a better condition than that out the shell stratum. Among the three sites, only the GB site was available for comparative testing, because all the skeletal remains in the SH and KT sites were excavated from the in-shell-stratum (Table 2). Four out of 14 samples were excavated from the in-shell-stratum and the others were from the out-shell-stratum in the GB site. We found, however, no significant difference between the in- and out-shell-stratum in MRs (Mann–Whitney U-test, P = 0.887) (Figure 2). Similarly, no significant association was found between the dating and the MRs (P = 0.656) (Figure 3).

Figure 2.

The association between the environmental factors (in/out shell stratum) and the MRs. Four individuals were excavated from the shell stratum; ten were excavated from the non-shell stratum. There was no significant difference between them (P = 0.887).

Figure 3.

The association between the environmental factors (dating) and the MRs in the three Jomon sites. The black line represents the regression line. There was no association between them (P = 0.651). Circles, triangles, and squares represent the Gionbara shell-mound site (n = 13), the Kikumatenaga shell-mound site (n = 15), and the Saihiro shell-mound site (n = 7), respectively.

Second, we tested an association between a molecular factor, CRR, and MR. The gelatin extracted from the 39 bone samples and the CRRs were calculated as a gelatin weight divided by a bone weight. Overall, there was no significant association between the 39 CRRs from all archaeological sites and their MRs (P = 0.091) (Figure 4a). When the association was examined for each site, no significant association was found between them (P = 0.160 for GB; P = 0.146 for SH; P = 0.939 for KT). Regarding the gelatin extraction for estimating the CRRs, we used different parts of bones for each sample (Table 1). The relationship between the CRRs and the MRs of different bone parts (ilium, limb, rib, and skull) is shown in Figure 4a. Among the four parts of bones, the statistical tests were performed only for skulls (n = 11) and for ribs (n = 22), since the numbers of samples were small for limb (n = 5) and ilium (n = 1). No significant association was found between the CRRs and the MRs for the skulls (P = 0.745) (Figure 4b), while a significant association was detected between the CRRs and MRs for ribs (P = 0.020) (Figure 4c). Thus, we found that the CRRs of ribs could be a strong candidate to be indicator for a good state of preservation of DNA in the skeletal remains.

Figure 4.

The association between the CRRs and the MRs. Circles, triangles, and squares represent the Gionbara shell-mound site, the Kikumatenaga shell-mound site, and the Saihiro shell-mound site, respectively, which include bone parts of limb, ilium, rib, and skull, as shown. The black line represents the regression line. (a) The CRRs and the MRs in the three Jomon sites are shown. There was no association between them (P = 0.090). (b) The CRRs and the MRs in the skulls (n = 11) from the three Jomon sites are shown. There was no association between them (P = 0.745). (c) The CRRs and the MRs in the ribs (n = 22) from the three Jomon sites are shown. There was an association between them (P = 0.020).

However, there was still a possibility that a spurious association existed due to our combining rib samples from three archaeological sites that have different preservation conditions. Therefore, we conducted an ANOVA to assess differences in the CRRs of ribs and the MRs between sites. We found the CRRs differed significantly among the three sites (P = 0.039 for CRRs; P = 0.318 for MRs), and therefore we assessed the relationship between the CRRs of ribs and MRs, taking into consideration differences between sites in the multiple regression analysis adjusted for site (i.e. sites were included in a linear model as dummy variables). The results revealed that the sites were not significantly associated with MRs, although the association between the CRRs and MRs of ribs was marginally significant (Supplementary Table 1). Multicollinearity was not shown for CRRs and sites (GVIF = 1.186 for CRRs; GVIF = 1.089 for sites). Therefore, we deduce that the correlation between the CRRs in rib bones and the MRs was statistically significant.

Discussion

In this study, we used MRs based on the reads from NGS, instead of the conventional PCR amplification, to assess the amount of endogenous DNA. As environmental factors, we tested in- or out-shell-stratum and the dating of the bones excavated from three archaeological sites (Figure 1, Table 2). We found no association between the environmental factors and the MRs (Figure 2, Figure 3). As a molecular factor, we looked at the CRR. When we tested all the samples combined together, we found no association between the CRRs and the MRs. Similarly, when we examined each archaeological site, we had no association between them (Figure 4a). Meanwhile, when we focused on the rib bones, we found a significant association between the CRRs and the MRs (Figure 4c). Although this might be an important finding, we still had to exclude the possibility of a spurious association caused by combining the rib samples from three archaeological sites with different preservation conditions. Therefore, we examined the relationship between the CRRs and MRs through regression analysis, adjusting for site-specific factors. The findings indicated that the individual archaeological sites did not exhibit a significant association with the MRs. However, it is worth noting that a marginal level of significance was observed in the relationship between MRs and CRRs in rib samples (Supplementary Table 1). In addition, we performed a single regression analysis to assess the relationship between CRRs of ribs and MRs for sites GB and KT, excluding SH due to its small sample size. We did not find any significant association between the CRRs of ribs and the MRs for each site (P = 0.144 for GB; P = 0.175 for KT), but there was a consistent trend in effect size (regression coefficient = 4.436 for GB; regression coefficient = 9.778 for KT). This appears to be because the sample size for each site was too small to detect a significant association in the regression analysis. Hence, we conclude that the CRRs in rib bones were significantly associated with the MRs.

The association between the environmental factors and DNA preservation is still unclear. The results shown in Figure 1 that there was no association in DNA survival between in- and out-shell-strata was unexpected, as soil pH should have a significant bearing on DNA preservation. Because the number of samples that could be compared in this study might be too small, additional analyses with a larger number of samples will be required in the future. On the other hand, as regards the lack of an association with calibrated radiocarbon dates shown in Figure 2, the result was to be expected. It is known that DNA is degraded soon after the death of the organism and the length of DNA (less than 100 bp in mtDNA) does not change for a long period (Sawyer et al., 2012). The degradation of DNA could be approximated as exponential decay (Allentoft et al., 2012). Our results are considered reasonable, as the samples are more than 3000 years postmortem, so a difference in age of several hundred years is not expected to affect DNA preservation.

Our study shows that a biomolecular factor, the CRR, looks very promising as an indicator of DNA preservation. We found that the CRR of the ribs in particular seems to be suitable as an indicator. Because ribs are flat bones whose thick spongy bone is sandwiched between two layers of thin compact bones, the degradation of collagen in rib bones could be faster than in other hard bones, such as skulls, which might be more parallel to the rate of DNA degradation. Collagen is always extracted for dating which must be done in ancient genome analyses, so no additional bone-crushing for calculating CRRs is involved. It is very attractive that the CRR can be calculated by Fourier-transform infrared spectroscopy (FTIR) without crushing bones (Naito et al., 2020). Thus, further research may lead to the development of FTIR-based technology to calculate CRRs that enables the measurement of DNA preservation without bone destruction.

Acknowledgments

This study was supported by JSPS KAKENHI Grant-in-Aid for Scientific Research (A) to 18H03590 and 22H00020 to R.T.

References

•  Allentoft  M.E.,  Collins  M.,  Harker  D.,  Haile  J.,  Oskam  C.L., et al (2012) The half-life of DNA in bone: measuring decay kinetics in 158 dated fossils. Proceedings of the Royal Society B: Biological Sciences, 279(1748): 4724–4733.
•  Bronk Ramsey  C. (2009) Bayesian analysis of radiocarbon dates. Radiocarbon, 51: 337–360.
•  Danecek  P.,  Bonfield  J.K.,  Liddle  J.,  Marshall  J.,  Ohan  V., et al (2021) Twelve years of SAMtools and BCFtools. Gigascience, 10: giab008.
•  DeNiro  M.J. (1985) Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature, 317: 806–809.
•  Fu  Q.,  Li  H.,  Moorjani  P.,  Jay  F.,  Slepchenko  S.M., et al (2014) Genome sequence of a 45,000-year-old modern human from Western Siberia. Nature, 514(7523): 445–449.
•  Gakuhari  T.,  Nakagome  S.,  Rasmussen  S.,  Allentoft  M.E.,  Sato  T., et al (2020) Ancient Jomon genome sequence analysis sheds light on migration patterns of early East Asian populations. Communications Biology, 3(1): 437.
•  Gamba  C.,  Jones  E.R.,  Teasdale  M.D.,  McLaughlin  R.L.,  Gonzalez-Fortes  G., et al (2014) Genome flux and stasis in a five millennium transect of European prehistory. Nature Communications, 5: 5257.
•  Gordon  C.C. and  Buikstra  J.E. (1981) Soil pH, bone preservation, and sampling bias at mortuary sites. American Antiquity, 46: 566–571.
•  Götherström  A.,  Collins  M.J.,  Angerbjörn  A., and  Lidén  K. (2002) Bone preservation and DNA amplification. Archaeometry, 44: 395–404.
•  Green  R.E.,  Krause  J.,  Ptak  S.E.,  Briggs  A.W.,  Ronan  M.T., et al (2006) Analysis of one million base pairs of Neanderthal DNA. Nature, 444: 330–336.
•  Green  R.E.,  Krause  J.,  Briggs  A.W.,  Kircher  M.,  Patterson  N., et al (2010) A draft sequence of the Neandertal genome. Science, 328(5979): 710–722.
•  Habu  J. (2004) Ancient Jomon of Japan. Cambridge University Press, Cambridge.
•  Hagelberg  E.,  Sykes  B., and  Hedges  R. (1989) Ancient bone DNA amplified. Nature, 342: 485.
•  Heaton  T.J.,  Köhler  P.,  Butzin  M.,  Bard  E.,  Reimer  R.W., et al (2020) Marine20—the marine radiocarbon age calibration curve (0–55,000 cal BP). Radiocarbon, 62: 779–820.
•  Hedges  R.E. and  Millard  A.R. (1995) Bones and groundwater: towards the modelling of diagenetic processes. Journal of Archaeological Science, 22: 155–164.
•  Jónsson  H.,  Ginolhac  A.,  Schubert  M.,  Johnson  P.L., and  Orlando  L. (2013) mapDamage2.0: fast approximate Bayesian estimates of ancient DNA damage parameters. Bioinformatics, 29: 1682–1684.
•  Kanzawa-Kiriyama  H.,  Kryukov  K.,  Jinam  T.A.,  Hosomichi  K.,  Saso  A., et al (2017) A partial nuclear genome of the Jomons who lived 3000 years ago in Fukushima, Japan. Journal of Human Genetics, 62: 213–221.
•  Kanzawa-Kiriyama  H.,  Jinam  T.A.,  Kawai  Y.,  Sato  T.,  Hosomichi  K., et al (2019) Late Jomon male and female genome sequences from the Funadomari site in Hokkaido, Japan. Anthropological Science, 127: 83–108.
•  Kondo  S. (1987) Jomon-jidai-syutsudo-jinkotsu to boiki. In: Kondo S. (ed.), Zaidan-houjin Ichiharashi Bunkazai Center Chousa-houkokusyo 23: Kikumatenaga-iseki. Zaidan-houjin Ichiharashi Bunkazai Center, Chiba, Japan, pp. 65–77 [in Japanese].
•  Krause  J.,  Briggs  A.W.,  Kircher  M.,  Maricic  T.,  Zwyns  N., et al (2010) A complete mtDNA genome of an early modern human from Kostenki, Russia. Current Biology, 20: 231–236.
•  Kurosaki  K.,  Matsushita  T., and  Ueda  S. (1993) Individual DNA identification from ancient human remains. American Journal of Human Genetics, 53: 638–643.
•  Lawlor  D.A.,  Dickel  C.D.,  Hauswirth  W.W., and  Parham  P. (1991) Ancient HLA genes from 7,500-year-old archaeological remains. Nature, 349(6312): 785–788.
•  Li  H. and  Durbin  R. (2009) Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics, 25: 1754–1760.
•  Lindahl  T. (1993) Instability and decay of the primary structure of DNA. Nature, 362(6422): 709–715.
•  Longin  R. (1971) New method of collagen extraction for radiocarbon dating. Nature, 230(5291): 241–242.
•  McColl  H.,  Racimo  F.,  Vinner  L.,  Demeter  F.,  Gakuhari  T., et al (2018) The prehistoric peopling of Southeast Asia. Science, 361(6397): 88–92.
•  Meyer  M.,  Kircher  M.,  Gansauge  M.T.,  Li  H.,  Racimo  F., et al (2012) A high-coverage genome sequence from an archaic Denisovan individual. Science, 338: 222–226.
•  Naito  Y.I.,  Yamane  M., and  Kitagawa  H. (2020) A protocol for using attenuated total reflection Fourier‐transform infrared spectroscopy for pre‐screening ancient bone collagen prior to radiocarbon dating. Rapid Communications in Mass Spectrometry, 34: e8720.
•  Oota  H.,  Saitou  N.,  Matsushita  T., and  Ueda  S. (1995) A genetic study of 2,000‐year‐old human remains from Japan using mitochondrial DNA sequences. American Journal of Physical Anthropology, 98: 133–145.
•  Oshizawa  N. (1999) Maisou-jinkotsu. In: Oshizawa N. (ed.), Kazusa Kokubunzidai Iseki-chousa-houkoku 5: Gionbara-kaizuka. Zaidan-houjin Ichiharashi Bunkazai Center and Ichiharashi kyouiku iinkai, Chiba, Japan, pp. 312–338 [in Japanese].
•  Pääbo  S.,  Gifford  J.A., and  Wilson  A.C. (1988) Mitochondrial DNA sequences from a 7000-year old brain. Nucleic Acids Research, 16: 9775–9787.
•  Pääbo  S.,  Higuchi  R.G., and  Wilson  A.C. (1989) Ancient DNA and the polymerase chain reaction: the emerging field of molecular archaeology. Journal of Biological Chemistry, 264: 9709–9712.
•  Pinhasi  R.,  Fernandes  D.,  Sirak  K.,  Novak  M.,  Connell  S., et al (2015) Optimal ancient DNA yields from the inner ear part of the human petrous bone. PloS One, 10: e0129102.
•  Poinar  H.N.,  Höss  M.,  Bada  J.L., and  Pääbo  S. (1996) Amino acid racemization and the preservation of ancient DNA. Science, 272(5263): 864–866.
•  Prüfer  K.,  Racimo  F.,  Patterson  N.,  Jay  F.,  Sankararaman  S., et al (2014) The complete genome sequence of a Neanderthal from the Altai Mountains. Nature, 505: 43–49.
•  Rasmussen  M.,  Li  Y.,  Lindgreen  S.,  Pedersen  J.S.,  Albrechtsen  A., et al (2010) Ancient human genome sequence of an extinct Palaeo-Eskimo. Nature, 463(7282): 757–762.
•  Reich  D.,  Green  R.E.,  Kircher  M.,  Krause  J.,  Patterson  N., et al (2010) Genetic history of an archaic hominin group from Denisova cave in Siberia. Nature, 468: 1053–1060.
•  Reimer  P.J.,  Austin  W.E.N.,  Bard  E.,  Bayliss  A.,  Blackwell  P.G., et al (2020) The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon, 62: 725–757.
•  Sakurai  A. (2005) Maiso-jinkotsu. In: Yasui K. (ed.), Ichiharashi-Saihiro-kaiduka2. Kazusa Kokubunzidai Iseki-chosa-houkoku center, Chiba, pp. 259–268 [in Japanese].
•  Sawyer  S.,  Krause  J.,  Guschanski  K.,  Savolainen  V., and  Pääbo  S. (2012) Temporal patterns of nucleotide misincorporations and DNA fragmentation in ancient DNA. PloS One, 7: e34131.
•  Schubert  M.,  Lindgreen  S., and  Orlando  L. (2016) AdapterRemoval v2: rapid adapter trimming, identification, and read merging. BMC Research Notes, 9: 88.
•  Shinoda  K.I. and  Kanai  S. (1999) Intracemetery genetic analysis at the Nakazuma Jomon site in Japan by mitochondrial DNA sequencing. Anthropological Science, 107: 129–140.
•  Stephan  E. (1997) Patterns of chemical change in fossil bones and various states of bone preservation associated with soil conditions. Anthropozoologica, 25: 173–180.
•  Yoneda  M.,  Hirota  M.,  Uchida  M.,  Tanaka  A.,  Shibata  Y., et al (2002) Radiocarbon and stable isotope analyses on the earliest Jomon skeletons from the Tochibara rockshelter, Nagano, Japan. Radiocarbon, 44: 549–557.
•  Yoshida  K.,  Hara  T.,  Kunikita  D.,  Miyazaki  Y.,  Sasaki  T., et al (2010) Pre-bomb marine reservoir ages in the Western Pacific. Radiocarbon, 52: 1197–1206.