The modifications that occur to the human body after death are complex and known to be affected by a variety of intrinsic and extrinsic factors. The rate of decomposition can vary significantly depending on the environment and even the manner of death. Nonetheless, the process of decomposition has been demonstrated to be predictable, providing opportunities to estimate the time elapsed since death (also known as postmortem interval [PMI]) based on gross morphological and/or microscopic changes to the body. Precise and accurate estimation of the PMI is crucial to help establishing the timeline of events surrounding death, can support medicolegal investigators with the identification of the deceased and can corroborate or negate other forensic evidence.

In the first hours after death, the body undergoes several postmortem changes, including progressive cooling (algor mortis), increased rigidity associated with muscle stiffness (rigor mortis), and pink-purplish discolouration, in light skinned individuals, caused by the lack of blood circulation and the settling of blood in the lowest areas (livor mortis) (Clark et al., 1997; Henssge and Madea, 2007; Madea, 2016). After these stages, as the time since death increases, the breaking down and liquefaction of the organs and other soft tissues will occur: a process referred to as putrefaction. The lack of oxygenated circulation induces cellular hypoxia, leading to swelling of the cells, and subsequent rupture of cell membranes and release of digestive enzymes. This triggers the autolytic digestion of soft tissues (Lee Goff, 2009). The body becomes fully anaerobic, allowing anoxic (endogenous) bacteria originated from the gut to proliferate and transmigrate throughout the entire body (Javan et al., 2016; Jans et al., 2004). The activity of endogenous bacteria results in the accumulation of gases which causes bloating of the soft tissues, starting from the abdomen but also taking place in the face during early decomposition stages, and progressing towards the rest of the body. Colonization of the body by insects and exogenous bacteria, mostly aerobic microorganisms, contributes further to the macroscopic changes and to the reduction of the overall mass of soft tissues (Hyde et al., 2013; Carter et al., 2017). Besides these, other extrinsic factors including abiotic environmental conditions (e.g., humidity, temperature, sun exposition, aeration, burial context) and biotic factors, such as the presence and type of microorganisms, insects, and scavengers (Cockle and Bell, 2015; Procopio et al., 2017b), will affect the rate of decomposition of the soft tissues. Intrinsic factors known to affect the rate of decomposition include, among others, body mass index and both antemortem and perimortem pathological conditions (Mickleburgh et al., 2021). Completion of the putrefactive stage and the activity of insects consuming the remaining soft tissues will leave the remains completely, or almost completely, skeletonized, and dry.

The complex nature and interplay of intrinsic and extrinsic variables involved in the process of decomposition makes the development of accurate and precise models for PMI estimation extremely challenging. Traditional methods of PMI estimation include calculating PMI using the body temperature and ambient temperature (which relies on the predictability of algor mortis, and works for short PMIs only), or the visual assessment of gross morphological changes to the body to estimate a relatively wide PMI range. Since the rate of gross morphological changes is variable, methods that rely on the visual scoring of decomposition stages suffer from issues of poor accuracy and precision. An additional problem of such methods is the effect of interobserver variations on the scoring of decomposition stages. For all commonly used PMI estimation methods, the accuracy and precision decreases considerably as decomposition progresses, and is particularly problematic when the remains are partially or completely skeletonized (Henssge and Madea, 2007; Madea, 2016).

In recent years, the number of studies exploring the use of biomolecular methods for PMI estimation has risen sharply, due to their potential for providing more accurate and precise estimation methods based on the rates of decay of different molecules and compounds (Procopio et al., 2018b; Prieto-Bonete et al., 2019; Pesko et al., 2020; Locci et al., 2019; Zelentsova et al., 2020). Better understanding of biomolecular decomposition of bone will provide opportunities to develop biomolecular methods for the estimation of longer PMIs (i.e., timeframes in which soft tissues are unlikely to be preserved). Moreover, through the combined analysis of multiple different panels of omics, greater precision and accuracy of PMI estimation can potentially be achieved.

Biomolecular decomposition is caused by both enzymatic and microbial breakdown of large molecules, resulting in the breakage of proteins into amino acids (AA), of carbohydrates into more simple monosaccharides, and of lipids into simpler fatty acids chains (Dent et al., 2004; Nolan et al., 2020). In carbohydrate decomposition, the complex polysaccharides are normally broken down via microbial activity into smaller units of monosaccharides. This breakdown can be fully achieved by oxidation, with the production of carbon dioxide and water, or can be partially achieved with the production of organic acids and alcohols. Alternatively, the monosaccharides can be degraded by fungal activity into glucuronic, citric, and oxalic acids, or by bacteria into lactic, butyric, and acetic acids (Dent et al., 2004; Stuart, 2013). During the decay of lipids, free saturated and unsaturated fatty acids are released due to the hydrolysis mediated by the action of intrinsic lipases released after death. These can then be converted into hydroxyl fatty acids (the main constituent of adipocere) by the action of specific bacterial enzymes in humid environments, or can associate with potassium and sodium ions, resulting in the formation of salts (Stuart, 2013). Protein degradation is primarily an enzyme-driven process, led by the action of proteases, which occurs at different rates for different proteins and tissues. Proteolytic enzymes induce the hydrolytic breakdown of proteins and the production respectively of proteoses, peptones, polypeptides, and finally AA, which can be further modified via deamination (production of ammonia), decarboxylation (production of cadaverine, putrescine, tyramine, tryptamine, indole, skatole, and carbon dioxide) and desulfhydralation (production of hydrogen sulphide, pyruvic acid, and thiols) (Dent et al., 2004; Stuart, 2013). Time-dependent non-enzymatic processes can also affect protein degradation and modifications (i.e.,deamidations).

The analysis of low molecular weight compounds and decomposition by-products is becoming more popular in forensic science, particularly for the purpose of estimating PMI (Locci et al., 2020). Time since death was recently reported as the main variable driving modifications in the metabolome occurring after death (Chighine et al., 2021) in many soft tissues and fluids, so the metabolomic approach appears ideal to estimate PMI. However, the potential forensic significance of the postmortem bone metabolome is as yet underexplored (Alldritt et al., 2019). Several studies on soft tissues (vitreous and aqueous humour) have examined metabolomics for the purpose of determining short PMIs. Examining longer PMIs based on metabolomics analysis of humour has not been possible due to evaporation and leakage through the corneal surface as time since death progresses (Locci et al., 2019). Girela et al., 2008 reported a significant positive correlation between PMI and taurine, glutamate, and aspartate levels found in vitreous humour. These results were partially confirmed by Zelentsova et al., 2020, who found a correlation between the levels of hypoxanthine, choline, creatine, betaine, glutamate, and glycine and PMI. Another approach employing ¹H-NMR on aqueous humour from pig heads reported taurine, choline, and succinate as major metabolites involved in the postmortem modification (Locci et al., 2019). The same study also showed an orthogonally constrained PLS2 model showing prediction error of 59 min for PMI <500 min, 104 min for PMI from 500 to 1000 min, and 118 min for PMI >1000 min. Besides humour, muscle is one of the most frequently targeted tissues in metabolomics studies focused on short PMI estimation. Pesko et al., 2020 recently evaluated rat and human biceps femoris muscles from the same individuals at different PMIs, demonstrating an increase of the abundance of several metabolites, including most of those derived from the breakdown of proteins, and in particular highlighting how threonine, tyrosine and lysine show the most consistent and predictable variations in relatively short PMIs. An untargeted metabolomics study on muscle tissue also indicated the potential of isolating biomarkers associated with age (Wilkinson et al., 2020), suggesting the potential applications of metabolomics for both age-at-death and PMI estimation.

To date, only three studies have used lipidomics assays for PMI estimation. Two of them were conducted on muscle tissue and showed, in general, a negative correlation between most lipid classes and PMI, as well as an increment in free fatty acids (Langley et al., 2019; Wood and Shirley, 2013). The third study applied lipidomics to trabecular bone samples from calcanei spanning a PMI of approximately 7 years and highlighted the presence of 76 potential N-acyl AA that could be employed for PMI estimation, however their correlation with PMI has not yet been fully elucidated (Dudzik et al., 2017).

Several studies have tried to quantify the degree of survival of proteins and the accumulation of post-translational modifications (PTMs) of AA in both animal and human models (Procopio et al., 2017b; Mickleburgh et al., 2021; Prieto-Bonete et al., 2019; Procopio et al., 2021; Mizukami et al., 2020) as well as under different conditions (e.g., in aquatic environments, different types of coffins, buried vs. surface) (Procopio et al., 2021; Mizukami et al., 2020; Bonicelli et al., 2022). The premise of these studies is that the protective action of the hydroxyapatite is expected to enhance the survival of proteins, allowing potential estimation of longer PMIs. Results generally showed that blood/plasma and ubiquitous proteins decrease in their abundance constantly starting from the early decomposition stages, whereas proteins more strongly connected to the mineral matrix such as bone-specific proteins are able to survive for longer PMIs and can be useful indicators for PMI estimation also in skeletonized remains. Similarly, also the accumulation of specific non-enzymatic PTMs, such as deamidations, can be used as a biomarker for the evaluation of the PMI in bones.

While many studies have applied different analytical platforms for proteomics, metabolomics, and lipidomics to several different matrices (Pesko et al., 2020; Locci et al., 2019; Zelentsova et al., 2020; Girela et al., 2008; Li et al., 2018; Wu et al., 2018; Hirakawa et al., 2009; Banaschak et al., 2005; Li et al., 2017), relatively little is known about the biomolecular decomposition of bone tissue. Moreover, while clinical studies have applied multi-omics methods with some frequency, their potential for the development of more precise and accurate biomolecular PMI estimation methods has not been explored. The present study applies, for the first time, a multi-omics approach (i.e., combined proteomics, metabolomics and lipidomics, defined here as the ‘ForensOMICS’ approach) to pre- and post-decomposition tibial cortical bone samples from four human female body donors, to identify potential multi-omics biomarkers of time since death. The multi-omics approach uses the natural differences in manner and rate of decomposition between the different biomolecules (proteins, metabolites, lipids) to expand the potential range of PMIs and to cross-correlate results between different sets of biomarkers to narrow down PMI ranges based on the degradation of multiple biomolecules. The use a of a single omics technique would not be suitable to investigate a wide range of potential PMIs. Metabolites and lipids are appropriate for short PMIs while protein has been proved to be stable across longer ones. Therefore, the combination of the three classes of biomolecules aims to obtain an ideal coverage across a wider range of PMIs. Additional advantages of the integration of different biomolecule classes might include greater flexibility in their application across different environments and different postmortem treatments, since it could increase the likelihood of retrieving suitable markers for PMI estimation. The present study provides a proof-of-concept for future validation of the multi-omics approach on a larger number of individuals.

This study comprises, to the best of our knowledge, the first attempt to apply a panel of three omics methods to human bones from a controlled decomposition experiment, to identify potential biomarkers for biomolecular PMI estimation. To develop and validate multi-omics PMI estimation methods for forensic applications, replication studies in substantial sample sizes of human bones will be necessary. However, the availability of bone samples both before and after decomposition from the same individuals is currently very limited. The work presented here represents a proof-of-concept study on the potential advantages of combining different omics for PMI estimation. The small number of individuals included is consistent with numbers generally used in human decomposition experiments, in which for practical and ethical reasons larger samples, such as used in clinical studies, are very difficult to obtain. While the sample size used here is not suitable for validation purposes, it serves to demonstrate the value and potential of the ‘ForensOMICS’ approach.

Considering each omics individually, the proteomic profile appears to show quite a considerable overlap between the individuals from three post-decomposition groups (i.e., 219, 790, and 834 days) suggesting that this method on its own does not provide sufficient sensitivity to segregate close PMIs (Figure 1—figure supplement 3). This could be due to the nature of these biomolecules; proteins, in fact, are highly stable and may be better suitable for long-term PMI estimation in forensic scenarios (Procopio et al., 2018b; Prieto-Bonete et al., 2019) as well as in the investigation of archaeological remains (Wadsworth et al., 2017; Warinner et al., 2022). Additionally, other analyses such as post-translational protein modifications may reveal a greater potential for PMI estimation in bones than the evaluation of the abundance of specific markers on their own (Procopio et al., 2018b). Employing a system biology approach for PMI estimation for forensic purposes by combining more than one class of biomolecules that have different postmortem stability (Dent et al., 2004) provides a more comprehensive biological explanation of the processes under investigation. This is achieved here by combining different layers of omics (i.e., metabolomics, lipidomics, and proteomics) to reconstruct the molecular profile of the overall system. The DIABLO model simultaneously identifies important markers to optimize the classification of a specific variable by combining multiple omics techniques (Singh et al., 2019). This is normally used to explain the biological mechanisms that determine a disease and its development, while in our case the main advantage is represented by the potential of selecting a pool of compounds that effectively explains, and could accurately estimate, PMI changes over an extended period of time. One interesting aspect of this approach is the difference in clustering between the metabolite and lipid blocks individually compared to the integration model. It can be seen in Figure 1—figure supplement 1 (metabolomics block) that samples with increasing PMIs seem to cluster further away from the pre-deposition sample in a time-dependent manner, with the 219 days PMI being closer to the fresh donors and the 872 days one being the furthest located. However, as suggested, the metabolomics profile of D2 seems to be significantly different from the other donors in the fresh state, and this could suggest that interindividual variation could affect the efficient clustering. This has been already highlighted in the proteomics work conducted on the same samples and was likely caused by the health condition of the donor prior to death (Mickleburgh et al., 2021). In contrast, the positioning of the PMI in the cluster tree behaves in the opposite way for lipids, where the various profiles seem not to be affected by any apparent interindividual variation in the fresh nor in the decomposed state (Figure 1—figure supplement 2). Considering now the clustering of the integrative model, it provides a clear classification of the PMIs obtained by the combination of the three single blocks. Since the approach chosen for this pilot study was discriminant analysis and PMI was provided to the model as a categorical variable, we believe that treating the response variable (PMI) as an ordinal or continuous variable on a larger sample size could improve the interpretation of the results and the forensic applicability of the methodology. Despite acknowledging these limitations, these preliminary results show the possibility of using multi-omics integration to identify different PMI groups. Furthermore, the results for proteomics, that individually does not allow discrimination for these specific time intervals, are integrated in the final model by retaining only the proteins that contribute to PMI identification.

Additionally, the presence of the two main clusters identified (fresh and skeletonized) has been driven by the greater differences between pre- and post-deposition. Conventionally, when performing method development for PMI estimation on bone samples collections, the baseline time is not available. Therefore, the differences captured with the analysis would be obtained on skeletonized samples only. We believe, however, that due to the uniqueness of the sample it was not ideal to remove the pre-deposition specimens. Despite these issues, we found moderate to high correlation between the omics blocks that allows their integration using the sparse algorithm (Singh et al., 2019) for PMI estimation.

Recently, literature has grown on the use of molecular studies via omics platforms, especially for short-term PMIs. Most of the studies involving metabolomics for PMI estimation focused on quickly degradable matrices (e.g., muscle, blood, humour) collected over a short period of time (<1 month) (Pesko et al., 2020; Locci et al., 2019; Banaschak et al., 2005; Ith et al., 2011; Ith et al., 2002). As previously mentioned, the analysis of proteins in bone has shown applicability to estimate relatively long PMIs in forensics (Procopio et al., 2018b; Mizukami et al., 2020; Procopio et al., 2018a) as well as to address archaeological questions (Ntasi et al., 2022; Pal Chowdhury et al., 2021; Brandt and Mannering, 2021; Richter et al., 2022; Brown et al., 2016), due to the prolonged survival of this type of biomolecules. Finally, according to the studies presented so far, it seems that postmortem changes of lipids could provide PMI estimation across several years, although there is great need for validation (Dudzik et al., 2017; Dudzik et al., 2020). The combination of these biomolecules’ classes in a multi-omics model could therefore be beneficial for estimating PMI across a broader range of potential PMIs. Metabolites and lipids offer accuracy in the short to medium term while proteins could be the main markers for longer PMIs due to their greater stability. Furthermore, variable selection (Singh et al., 2019; Rohart et al., 2017) would offer the advantage of simplifying experimental procedures and targets those markers that behave consistently with PMI. To limit the potential effects of interindividual variability, we considered variables that showed no outliers among the four body donors and created a model that limits as much as possible the number of predictors without affecting the assessment of the PMI.

Our results for the metabolomics assay display clear differences between the pre- and post-placement bone metabolomic profiles, suggesting the potential to use these profiles to assess long PMIs. The small sample size in this study does not allow us to make any deep inferences about the biological significance of the metabolomics profiles of the post-placement samples, as these may have been influenced by exogenous factors. With regard to the pre-placement samples, the PMIs ranging between 2 and 10 days at 4°C would have allowed some minimal postmortem modifications in the metabolome to occur (Chighine et al., 2021). The metabolomic profiles of these samples are characterized by creatine, taurine, hypoxanthine, 3-hydroxybutyrate, creatinine, and phenylaniline. Hypoxanthine is a well-known hallmark of ATP consumption and, consequently, a sign of exhaustion of normal substrates (i.e., glucose and pyruvate) of the tri-carboxylic acid (TCA) cycle. In conjunction with the presence of creatine, taurine, creatinine, phenylalanine, and 3-hydroxybutyrate, we may hypothesize a switch towards TCA cycle anaplerosis through amino acidic and ketonic substrates, in pursuit of a resilient ATP production during the early/mid-PMIs. Not only was the proposed metabolomic approach able to identify the pre- and post-deposition groups according to the bone metabolome modifications, but it was also sensitive enough to detect at very long PMIs. The presence of exogenous compounds (i.e., caffeine, ecgonine, dextromethorphan, tramadol N-oxide, penbutolol, salicylic acid) that could reflect lifestyle habits or pharmacological therapies, and thus potentially has major implications in forensic toxicology and personal identification, is consistent with evidence from animal models (Alldritt et al., 2019). Enrichment analysis can be found in Figure 3.

Figure 3

Download asset Open asset

Several polar metabolites identified in this study have previously been found in other tissues to show a consistent decay pattern after death. In fact, most of the compounds of interest matched here have already been flagged in other tissues as good potential biomarkers of PMI across shorter timeframes (Figure 2A). Uracil, a pyrimidine base of RNA, was previously seen to increase over a 14-day PMI in human muscle tissue when analysed by LC-MS (Pesko et al., 2020). Similar results for this compound were found in GC-MS analysis of rat’s blood (Dai et al., 2019). In contrast, no clear association between this metabolite and PMI was found in aqueous humour (Locci et al., 2019). In the present study, after a drop in normalized intestines between the baseline and first PMI, we detected an increase until 834 days, and a drop towards the longest PMI considered. It is worth mentioning that most metabolites drop significantly after the baseline (‘fresh’) times (Figure 2A), suggesting that compound decomposition is driving this first part of the PMI following the stop of human metabolism. It is interesting that with the increase in PMI there is also an increment in several compounds that could be associated with the breakdown of larger biomolecules (e.g., proteins) or with the presence of microbial communities that leave their own metabolic profile on bone surface. Another common marker of interest is hypoxanthine for its association with hypoxia (Locci et al., 2019; Zelentsova et al., 2020; Kaszynski et al., 2016; Jawor et al., 2019; Locci et al., 2021), which seems to drastically drop between the baseline times and the first PMI timepoint, as well as in the last time interval, showing a good consistency with PMI. In contrast, hypoxanthine was seen to increase until 48 hr and then to decrease at 72 hr in rat blood (Donaldson and Lamont, 2013). Zelentsova et al., 2020, showed a positive relation between hypoxanthine and PMI in human serum, aqueous and vitreous humour. To fully understand the behaviour of this compound in bone tissue, a longitudinal study should be performed, also including short PMIs. Leucine has also been reported in short time scale to increase in human muscle tissue (Pesko et al., 2020) and this agrees with our results where, after the initial drop, we noticed a consistent increase from the first PMI onwards. What can be clearly seen in Figure 2A is that D2 affects the linearity of the trend, suggesting that there might be some degree of interindividual variability. This is the case for several compounds; this limitation could be mitigated by increasing the number of individuals per timepoint in future studies. Creatinine has previously been reported to be a good marker in both muscle tissue (Pesko et al., 2020). Although it has not been mentioned in literature previously, we also found that neopterin, a biomarker for immune system activation commonly profiled in blood, serum, and urine (Melichar et al., 2017; Laudanski et al., 2021), has a strong negative correlation with PMI. Taurine, also in accordance with studies on vitreous humour (Locci et al., 2019), showed a predictable positive behaviour with PMI. Acetamide is a nitrogen-based compound associated with active and advanced decay (Dekeirsschieter et al., 2012) that, not surprisingly, showed the best positive association with PMI, resulting in being the most reliable biomarker within the entire panel considered. Palmitoylethanolamide is a carboximidic acid that was shown to accumulate in relation with cellular stress in pig brains postmortem (Buczynski and Parsons, 2010). These findings agree with our study, which revealed a clear increase of this metabolite with increasing PMIs. N,N-diethylethanolamine, belonging to the class of organic compounds known as 1,2-aminoalcohols, has not yet been highlighted for its potential in PMI estimation. In the current study, this molecule increased in the decomposed samples, although no clear trends were observed across the various PMIs. A proposed mechanism for its accumulation is the partial oxidation driven by bacterial decomposition of monosaccharides into organic alcohols (Dent et al., 2004; Nolan et al., 2020). 12-Aminododecanoic acid and 12-hydroxydodecanoic acid are instead medium-chain fatty acids that show a positive relationship with PMI. Previous studies based on skeletal muscle tissue reported a decline in very-long-chain fatty acids (Langley et al., 2019; Wood and Shirley, 2013) in very short PMIs. It is not possible to exclude the cleavage of longer chains by the action of lipases or microorganic activity (Dent et al., 2004; Stuart, 2013). The last compound selected in the final model is methylmalonic acid, a carboxylic acid which is an intermediate in the metabolism of fat and proteins. It has been shown that abnormally high levels of organic acids in blood (organic acidaemia), urine (organic aciduria), brain, and other tissues lead to general metabolic acidosis (Narayanan et al., 2011). In this study, even with a postmortem increase in its concentration, it is not possible to identify a clear trend across the decomposed samples; this may be related to interindividual biological differences of the donors involved in this study (e.g., age and health condition).

From the lipidomic assay, only five markers were selected in the final model. These are three lysophosphatidylcholines (LPCs), one phosphatidylcholine (PC) and one phosphatidylinositol (PI), all showing decreasing intensities in the decomposed samples in comparison with the ‘fresh’ ones. PCs are generally the most abundant neutral phospholipids and represent the main constituent in cellular membranes. LPCs are derived from the hydrolysis of dietary and biliary PCs and are absorbed as such in the intestines, but they become re-esterified before being exported in the lymph (McMaster, 2018). They are present in cell membranes and in blood. Their half-life in vivo is limited because of the quick metabolic reaction that involves lysophospholipases and LPC-acyltransferases (Law et al., 2019). In contrast, PLS are amphiphilic molecules that are also minorly present in cell membranes, whose role is to modulate the membrane curvature and to have other bioactive functions such as interacting with peripheral proteins (Falkenburger et al., 2010) and inhibiting osteoclast formation (Alhouayek et al., 2018). After death, these compounds can be converted into fatty acids via hydrolysis to then hydrogenize or oxidase to form saturated and unsaturated fatty acids (Dent et al., 2004). This process is driven by intrinsic tissues lipases (Dent et al., 2004). A very limited number of studies have applied lipidomics for PMI estimation. Langley et al., 2019, evaluated human skeletal muscle tissue from 31 donors over a PMI of 2000 accumulated degree days showing consistent extraction of phosphatidylglycerol (PG) 34:0 and phosphatidylethanolamine 36:4, which showed good correlation with PMI. Wood and Shirley, 2013, investigated the lipidome of human anterior quadriceps muscle from one donor at 1-, 9-, and 24-day PMIs showing the decline of sterol sulphates, choline plasmalogens, ethanolamine plasmalogens, and PGs and the increase of free fatty acids. Our results lend support to these earlier findings and further confirm the potential of lipidomics for PMI estimation. Nonetheless, direct comparison with these studies is not possible as they considered different tissues for much shorter PMIs. Additionally, lipids profiled from the muscle tissue after decomposition are suggested to derive from cell membrane breakdown (Langley et al., 2019; Wood and Shirley, 2013). We suggest that, in bone material, the lipidome under investigation accounts not only for cell membrane decomposition of embedded osteocytes but also for the marrow and fluids embedded in the bone pores.

The proteomics results revealed that two ubiquitous proteins (histones), haemoglobin, ACTB, and VIME are the best candidates within this multi-omics PMI model. These five proteins selected by the model represent those which were best able to discriminate between the ‘fresh’ bones and the ‘skeletonized’ bones but are therefore not necessarily the best biomarkers to differentiate between the four post-decomposition PMIs. For insights on the most suitable protein biomarkers for differentiating between the longer PMIs, identified by excluding the ‘fresh’ samples, see Mickleburgh et al., 2021. It is not surprising to see that the proteins highlighted in the model are either ubiquitous proteins or blood or muscle tissue proteins, as their abundance would naturally be higher in ‘fresh’ bone than in ‘skeletonized’ bones. The HBA is found in red blood cells but is often also identified in bone samples with long PMIs from archaeological contexts (Smith and Wilson, 1990), and its consistent time-dependent degradation has been previously highlighted in skeletal remains using several platforms (Ramsthaler et al., 2011; Wiley et al., 2009). Furthermore, it has already been reported in skeletal tissue from controlled decomposition studies of animals, and already highlighted as a potential biomarker for PMI estimation (Procopio et al., 2018b). VIME was also previously reported by Procopio et al., 2018b, to be associated with PMI. It is a filament protein abundant in muscle tissue, and therefore its association with bone, particularly with the ‘fresh’ samples, is not unexpected. However, we emphasize that this could also be due to interindividual variability, and that further investigation may clarify the usefulness of VIME to estimate PMI. ACTB, similar to VIME, is a structural protein that forms cross-linked networks in the cytoplasmatic compartments and that is strongly connected with the presence of muscle tissue residues. A previous study showed the decrease in myosin contents with increasing PMIs, similarly to what we observed here for ACTB. The remaining two proteins are both components of the nucleosomes, in our study were shown to be drastically reduced in bone tissue also at the first the baseline PMI taken into consideration. In sum, these results allowed the identification of five protein biomarkers which make good candidates for estimation of short PMIs (<900 days) (e.g., considering time points limited to months postmortem) and not for years after death for which structural and functional proteins in bone have been shown better targets to employ for PMI estimation (Mickleburgh et al., 2021; Prieto-Bonete et al., 2019).

Based on the findings of this exploratory study, we argue that the multi-omic method we adopted here shows considerable potential for the future development of an accurate and precise PMI estimation method for human bone. Further research should focus on increasing the sample size, to ultimately validate the method for application in forensic investigation of skeletonized human remains. Beyond the findings discussed at length above, we emphasize that it is of paramount importance to establish which biomolecules identified here are associated with the human metabolism and degradation, and which are produced by the decomposers’ microbial activity. Controlled taphonomic experiments on human decomposition at human taphonomy facilities provide the opportunity to elucidate biomolecular decomposition of human bone. A comprehensive understanding of the origin of different compounds is key to provide a detailed explanation of the postmortem changes that affect bone and other tissues, ultimately helping to shed a light on biomolecular PMI investigations and on the real potential that multi-omics analyses can have in this direction.

This data is available at the NIH Common Fund's National Metabolomics Data Repository (NMDR) website, the Metabolomics Workbench, with Study ID ST002283. The data can be accessed via Project DOI https://doi.org/10.21228/M8MH6X. The R script used is available in Source code 1.

1. 1. Bonicelli A

   (2022) Metabolomics Workbench

   ForensOMICS.

   https://doi.org/10.21228/M8MH6X

1. 1. Alhouayek M
   2. Masquelier J
   3. Muccioli GG

   (2018) Lysophosphatidylinositols, from cell membrane constituents to GPR55 ligands

   Trends in Pharmacological Sciences 39:586–604.

   https://doi.org/10.1016/j.tips.2018.02.011

   • PubMed
   • Google Scholar
2. 1. Alldritt I
   2. Whitham-Agut B
   3. Sipin M
   4. Studholme J
   5. Trentacoste A
   6. Tripp JA
   7. Cappai MG
   8. Ditchfield P
   9. Devièse T
   10. Hedges REM
   11. McCullagh JSO

   (2019) Metabolomics reveals diet-derived plant polyphenols accumulate in physiological bone

   Scientific Reports 9:8047.

   https://doi.org/10.1038/s41598-019-44390-1

   • PubMed
   • Google Scholar
3. 1. Banaschak S
   2. Rzanny R
   3. Reichenbach JR
   4. Kaiser WA
   5. Klein A

   (2005) Estimation of postmortem metabolic changes in porcine brain tissue using 1H-MR spectroscopy -- preliminary results

   International Journal of Legal Medicine 119:77–79.

   https://doi.org/10.1007/s00414-004-0495-7

   • PubMed
   • Google Scholar
4. 1. Bonicelli A
   2. Di Nunzio A
   3. Di Nunzio C
   4. Procopio N

   (2022) Insights into the differential preservation of bone proteomes in inhumed and entombed cadavers from Italian forensic caseworks

   Journal of Proteome Research 21:1285–1298.

   https://doi.org/10.1021/acs.jproteome.1c00904

   • PubMed
   • Google Scholar
5. 1. Brandt LØ
   2. Mannering U

   (2021) Taxonomic identification of danish viking age shoes and skin objects by zooms (zooarchaeology by mass spectrometry)

   Journal of Proteomics 231:104038.

   https://doi.org/10.1016/j.jprot.2020.104038

   • PubMed
   • Google Scholar
6. 1. Brown S
   2. Higham T
   3. Slon V
   4. Pääbo S
   5. Meyer M
   6. Douka K
   7. Brock F
   8. Comeskey D
   9. Procopio N
   10. Shunkov M
   11. Derevianko A
   12. Buckley M

   (2016) Identification of a new hominin bone from denisova cave, siberia using collagen fingerprinting and mitochondrial DNA analysis

   Scientific Reports 6:23559.

   https://doi.org/10.1038/srep23559

   • PubMed
   • Google Scholar
7. 1. Buczynski MW
   2. Parsons LH

   (2010) Quantification of brain endocannabinoid levels: methods, interpretations and pitfalls

   British Journal of Pharmacology 160:423–442.

   https://doi.org/10.1111/j.1476-5381.2010.00787.x

   • PubMed
   • Google Scholar
8. Book

   1. Carter DO
   2. Tomberlin JK
   3. Benbow ME
   4. Metcalf JL

   (2017) Microbial Communities Associated with Decomposing Corpses

   John Wiley & Sons.

   https://doi.org/10.1002/9781119062585

   • Google Scholar
9. 1. Chighine A
   2. Locci E
   3. Nioi M
   4. d’Aloja E

   (2021) Looking for post-mortem metabolomic standardization: waiting for godot-the importance of post-mortem interval in forensic metabolomics

   Chemical Research in Toxicology 34:1946–1947.

   https://doi.org/10.1021/acs.chemrestox.1c00211

   • PubMed
   • Google Scholar
10. Book

    1. Clark MA
    2. Worrell MB
    3. Pless JE

    (1997)

    Postmortem changes in soft tissues

    In: Haglund WD, Sorg MH, editors. Forensic Taphonomy: The Postmortem Fate of Human Remains. CRC Press. pp. 156–164.

    • Google Scholar
11. 1. Cockle DL
    2. Bell LS

    (2015) Human decomposition and the reliability of a “universal” model for post mortem interval estimations

    Forensic Science International 253:136.

    https://doi.org/10.1016/j.forsciint.2015.05.018

    • PubMed
    • Google Scholar
12. 1. Dai X
    2. Fan F
    3. Ye Y
    4. Lu X
    5. Chen F
    6. Wu Z
    7. Liao L

    (2019) An experimental study on investigating the postmortem interval in dichlorvos poisoned rats by GC/MS-based metabolomics

    Legal Medicine 36:28–36.

    https://doi.org/10.1016/j.legalmed.2018.10.002

    • PubMed
    • Google Scholar
13. 1. Dekeirsschieter J
    2. Stefanuto PH
    3. Brasseur C
    4. Haubruge E
    5. Focant JF

    (2012) Enhanced characterization of the smell of death by comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry (gcxgc-TOFMS)

    PLOS ONE 7:e39005.

    https://doi.org/10.1371/journal.pone.0039005

    • PubMed
    • Google Scholar
14. 1. Dent BB
    2. Forbes SL
    3. Stuart BH

    (2004) Review of human decomposition processes in soil

    Environmental Geology 45:576–585.

    https://doi.org/10.1007/s00254-003-0913-z

    • Google Scholar
15. 1. Donaldson AE
    2. Lamont IL

    (2013) Biochemistry changes that occur after death: potential markers for determining post-mortem interval

    PLOS ONE 8:e82011.

    https://doi.org/10.1371/journal.pone.0082011

    • PubMed
    • Google Scholar
16. Conference

    1. Dudzik B
    2. Langley NR
    3. Meadows Jantz L
    4. Cebak J
    5. Wood P

    (2017)

    Postmortem Interval Estimation Using Bone Lipidomics

    in American Academy of Forensic Sciences 70th Annual Scientific Meeting.

    • Google Scholar
17. Report

    1. Dudzik B
    2. Jantz LM
    3. Langley NR
    4. Wood P

    (2020)

    Postmortem Interval Determination from Bone: A Metabolomics and Lipidomics Approach

    NIJ report 304294.

    • Google Scholar
18. 1. Falkenburger BH
    2. Jensen JB
    3. Dickson EJ
    4. Suh BC
    5. Hille B

    (2010) Phosphoinositides: lipid regulators of membrane proteins modulation of K + and ca 2+ channels by PIP 2. he discovered PIP 2 regulation of KCNQ channels

    The Authors. Journal Compilation C 588:3179–3185.

    https://doi.org/10.1113/jphysiol.2010.192153

    • Google Scholar
19. 1. Folch J
    2. Lees M
    3. Sloane stanley GH

    (1957) A simple method for the isolation and purification of total lipides from animal tissues

    The Journal of Biological Chemistry 226:497–509.

    https://doi.org/10.1016/S0021-9258(18)64849-5

    • PubMed
    • Google Scholar
20. 1. Girela E
    2. Villanueva E
    3. Irigoyen P
    4. Girela V
    5. Hernández-Cueto C
    6. Peinado JM

    (2008) Free amino acid concentrations in vitreous humor and cerebrospinal fluid in relation to the cause of death and postmortem interval

    Journal of Forensic Sciences 53:730–733.

    https://doi.org/10.1111/j.1556-4029.2008.00726.x

    • PubMed
    • Google Scholar
21. Website

    1. Health and Safety Code

    (2009) Death and Disposition of the Body, Disposition of the Body, Revised Uniform Anatomical Gift Act

    Accessed March 20, 2022.

    https://statutes.capitol.texas.gov/Docs/HS/htm/HS.692A.htm
22. 1. Henssge C
    2. Madea B

    (2007) Estimation of the time since death

    Forensic Science International 165:182–184.

    https://doi.org/10.1016/j.forsciint.2006.05.017

    • PubMed
    • Google Scholar
23. 1. Hervé MR
    2. Nicolè F
    3. Lê Cao KA

    (2018) Multivariate analysis of multiple datasets: a practical guide for chemical ecology

    Journal of Chemical Ecology 44:215–234.

    https://doi.org/10.1007/s10886-018-0932-6

    • PubMed
    • Google Scholar
24. 1. Hirakawa K
    2. Koike K
    3. Uekusa K
    4. Nihira M
    5. Yuta K
    6. Ohno Y

    (2009) Experimental estimation of postmortem interval using multivariate analysis of proton NMR metabolomic data

    Legal Medicine 11 Suppl 1:S282–S285.

    https://doi.org/10.1016/j.legalmed.2009.02.007

    • PubMed
    • Google Scholar
25. 1. Hyde ER
    2. Haarmann DP
    3. Lynne AM
    4. Bucheli SR
    5. Petrosino JF

    (2013) The living dead: bacterial community structure of a cadaver at the onset and end of the bloat stage of decomposition

    PLOS ONE 8:e77733.

    https://doi.org/10.1371/journal.pone.0077733

    • PubMed
    • Google Scholar
26. 1. Ith M
    2. Bigler P
    3. Scheurer E
    4. Kreis R
    5. Hofmann L
    6. Dirnhofer R
    7. Boesch C

    (2002) Observation and identification of metabolites emerging during postmortem decomposition of brain tissue by means of in situ 1H-magnetic resonance spectroscopy

    Magnetic Resonance in Medicine 48:915–920.

    https://doi.org/10.1002/mrm.10294

    • PubMed
    • Google Scholar
27. 1. Ith M
    2. Scheurer E
    3. Kreis R
    4. Thali M
    5. Dirnhofer R
    6. Boesch C

    (2011) Estimation of the postmortem interval by means of ¹H MRS of decomposing brain tissue: influence of ambient temperature

    NMR in Biomedicine 24:791–798.

    https://doi.org/10.1002/nbm.1623

    • PubMed
    • Google Scholar
28. 1. Jans MME
    2. Nielsen-Marsh CM
    3. Smith CI
    4. Collins MJ
    5. Kars H

    (2004) Characterisation of microbial attack on archaeological bone

    Journal of Archaeological Science 31:87–95.

    https://doi.org/10.1016/j.jas.2003.07.007

    • Google Scholar
29. 1. Javan GT
    2. Finley SJ
    3. Can I
    4. Wilkinson JE
    5. Hanson JD
    6. Tarone AM

    (2016) Human thanatomicrobiome succession and time since death

    Scientific Reports 6:29598.

    https://doi.org/10.1038/srep29598

    • PubMed
    • Google Scholar
30. 1. Jawor P
    2. Ząbek A
    3. Wojtowicz W
    4. Król D
    5. Stefaniak T
    6. Młynarz P

    (2019) Metabolomic studies as a tool for determining the post-mortem interval (PMI) in stillborn calves

    BMC Veterinary Research 15:189.

    https://doi.org/10.1186/s12917-019-1935-4

    • PubMed
    • Google Scholar
31. 1. Kaszynski RH
    2. Nishiumi S
    3. Azuma T
    4. Yoshida M
    5. Kondo T
    6. Takahashi M
    7. Asano M
    8. Ueno Y

    (2016) Postmortem interval estimation: a novel approach utilizing gas chromatography/mass spectrometry-based biochemical profiling

    Analytical and Bioanalytical Chemistry 408:3103–3112.

    https://doi.org/10.1007/s00216-016-9355-9

    • PubMed
    • Google Scholar
32. 1. Langley N
    2. Wood P
    3. Herring P
    4. Steadman D

    (2019) Forensic postmortem interval estimation from skeletal muscle tissue: a lipidomics approach

    Forensic Anthropology 2:1011.

    https://doi.org/10.5744/fa.2019.1011

    • Google Scholar
33. 1. Laudanski K
    2. Okeke T
    3. Hajj J
    4. Siddiq K
    5. Rader DJ
    6. Wu J
    7. Susztak K

    (2021) Longitudinal urinary biomarkers of immunological activation in covid-19 patients without clinically apparent kidney disease versus acute and chronic failure

    Scientific Reports 11:19675.

    https://doi.org/10.1038/s41598-021-99102-5

    • PubMed
    • Google Scholar
34. 1. Law S-H
    2. Chan M-L
    3. Marathe GK
    4. Parveen F
    5. Chen C-H
    6. Ke L-Y

    (2019) An updated review of lysophosphatidylcholine metabolism in human diseases

    International Journal of Molecular Sciences 20:1149.

    https://doi.org/10.3390/ijms20051149

    • PubMed
    • Google Scholar
35. 1. Lee Goff M

    (2009) Early post-mortem changes and stages of decomposition in exposed cadavers

    Experimental & Applied Acarology 49:21–36.

    https://doi.org/10.1007/s10493-009-9284-9

    • PubMed
    • Google Scholar
36. 1. Li C
    2. Li Z
    3. Tuo Y
    4. Ma D
    5. Shi Y
    6. Zhang Q
    7. Zhuo X
    8. Deng K
    9. Chen Y
    10. Wang Z
    11. Huang P

    (2017) MALDI-tof MS as a novel tool for the estimation of postmortem interval in liver tissue samples

    Scientific Reports 7:4887.

    https://doi.org/10.1038/s41598-017-05216-0

    • PubMed
    • Google Scholar
37. 1. Li W
    2. Chang Y
    3. Cheng Z
    4. Ling J
    5. Han L
    6. Li X
    7. Ding Y

    (2018) Vitreous humor: a review of biochemical constituents in postmortem interval estimation

    Journal of Forensic Science and Medicine 4:85.

    https://doi.org/10.4103/jfsm.jfsm_13_18

    • Google Scholar
38. 1. Locci E
    2. Stocchero M
    3. Noto A
    4. Chighine A
    5. Natali L
    6. Napoli PE
    7. Caria R
    8. De-Giorgio F
    9. Nioi M
    10. d’Aloja E

    (2019) A 1H NMR metabolomic approach for the estimation of the time since death using aqueous humour: an animal model

    Metabolomics 15:76.

    https://doi.org/10.1007/s11306-019-1533-2

    • PubMed
    • Google Scholar
39. 1. Locci E
    2. Bazzano G
    3. Chighine A
    4. Locco F
    5. Ferraro E
    6. Demontis R
    7. d’Aloja E

    (2020) Forensic NMR metabolomics: one more arrow in the quiver

    Metabolomics 16:118.

    https://doi.org/10.1007/s11306-020-01743-6

    • PubMed
    • Google Scholar
40. 1. Locci E
    2. Stocchero M
    3. Gottardo R
    4. De-Giorgio F
    5. Demontis R
    6. Nioi M
    7. Chighine A
    8. Tagliaro F
    9. d’Aloja E

    (2021) Comparative use of aqueous humour 1H NMR metabolomics and potassium concentration for PMI estimation in an animal model

    International Journal of Legal Medicine 135:845–852.

    https://doi.org/10.1007/s00414-020-02468-w

    • PubMed
    • Google Scholar
41. 1. Madea B

    (2016) Methods for determining time of death

    Forensic Science, Medicine, and Pathology 12:451–485.

    https://doi.org/10.1007/s12024-016-9776-y

    • PubMed
    • Google Scholar
42. 1. McMaster CR

    (2018) From yeast to humans-roles of the Kennedy pathway for phosphatidylcholine synthesis

    FEBS Letters 592:1256–1272.

    https://doi.org/10.1002/1873-3468.12919

    • PubMed
    • Google Scholar
43. 1. Megyesi MS
    2. Nawrocki SP
    3. Haskell NH

    (2005) Using accumulated degree-days to estimate the postmortem interval from decomposed human remains

    Journal of Forensic Sciences 50:618–626.

    https://doi.org/10.1520/JFS2004017

    • PubMed
    • Google Scholar
44. 1. Melichar B
    2. Spisarová M
    3. Bartoušková M
    4. Krčmová LK
    5. Javorská L
    6. Študentová H

    (2017) Neopterin as a biomarker of immune response in cancer patients

    Annals of Translational Medicine 5:280.

    https://doi.org/10.21037/atm.2017.06.29

    • PubMed
    • Google Scholar
45. 1. Mickleburgh HL
    2. Schwalbe EC
    3. Bonicelli A
    4. Mizukami H
    5. Sellitto F
    6. Starace S
    7. Wescott DJ
    8. Carter DO
    9. Procopio N

    (2021) Human bone proteomes before and after decomposition: investigating the effects of biological variation and taphonomic alteration on bone protein profiles and the implications for forensic proteomics

    Journal of Proteome Research 20:2533–2546.

    https://doi.org/10.1021/acs.jproteome.0c00992

    • PubMed
    • Google Scholar
46. 1. Mizukami H
    2. Hathway B
    3. Procopio N

    (2020) Aquatic decomposition of mammalian corpses: a forensic proteomic approach

    Journal of Proteome Research 19:2122–2135.

    https://doi.org/10.1021/acs.jproteome.0c00060

    • PubMed
    • Google Scholar
47. 1. Narayanan MP
    2. Kannan V
    3. Vinayan KP
    4. Vasudevan DM

    (2011) Diagnosis of major organic acidurias in children: two years experience at a tertiary care centre

    Indian Journal of Clinical Biochemistry 26:347–353.

    https://doi.org/10.1007/s12291-011-0111-9

    • PubMed
    • Google Scholar
48. 1. Nolan AN
    2. Mead RJ
    3. Maker G
    4. Speers SJ

    (2020) A review of the biochemical products produced during mammalian decomposition with the purpose of determining the post-mortem interval

    Australian Journal of Forensic Sciences 52:477–488.

    https://doi.org/10.1080/00450618.2019.1589571

    • Google Scholar
49. 1. Ntasi G
    2. Palomo IR
    3. Marino G
    4. Piaz FD
    5. Cappellini E
    6. Birolo L
    7. Petrone P

    (2022) Molecular signatures written in bone proteins of 79 AD victims from herculaneum and pompeii

    Scientific Reports 12:8401.

    https://doi.org/10.1038/s41598-022-12042-6

    • PubMed
    • Google Scholar
50. 1. Pal Chowdhury M
    2. Choudhury KD
    3. Bouchard GP
    4. Riel-Salvatore J
    5. Negrino F
    6. Benazzi S
    7. Slimak L
    8. Frasier B
    9. Szabo V
    10. Harrison R
    11. Hambrecht G
    12. Kitchener AC
    13. Wogelius RA
    14. Buckley M

    (2021) Machine learning ATR-FTIR spectroscopy data for the screening of collagen for zooms analysis and mtdna in archaeological bone

    Journal of Archaeological Science 126:105311.

    https://doi.org/10.1016/j.jas.2020.105311

    • Google Scholar
51. 1. Pesko BK
    2. Weidt S
    3. McLaughlin M
    4. Wescott DJ
    5. Torrance H
    6. Burgess K
    7. Burchmore R

    (2020) Postmortomics: the potential of untargeted metabolomics to highlight markers for time since death

    OMICS 24:649–659.

    https://doi.org/10.1089/omi.2020.0084

    • PubMed
    • Google Scholar
52. 1. Prieto-Bonete G
    2. Pérez-Cárceles MD
    3. Maurandi-López A
    4. Pérez-Martínez C
    5. Luna A

    (2019) Association between protein profile and postmortem interval in human bone remains

    Journal of Proteomics 192:54–63.

    https://doi.org/10.1016/j.jprot.2018.08.008

    • PubMed
    • Google Scholar
53. 1. Procopio N
    2. Buckley M

    (2017a) Minimizing laboratory-induced decay in bone proteomics

    Journal of Proteome Research 16:447–458.

    https://doi.org/10.1021/acs.jproteome.6b00564

    • PubMed
    • Google Scholar
54. 1. Procopio N
    2. Chamberlain AT
    3. Buckley M

    (2017b) Intra- and interskeletal proteome variations in fresh and buried bones

    Journal of Proteome Research 16:2016–2029.

    https://doi.org/10.1021/acs.jproteome.6b01070

    • PubMed
    • Google Scholar
55. 1. Procopio N
    2. Chamberlain AT
    3. Buckley M

    (2018a) Exploring biological and geological age-related changes through variations in intra- and intertooth proteomes of ancient dentine

    Journal of Proteome Research 17:1000–1013.

    https://doi.org/10.1021/acs.jproteome.7b00648

    • PubMed
    • Google Scholar
56. 1. Procopio N
    2. Williams A
    3. Chamberlain AT
    4. Buckley M

    (2018b) Forensic proteomics for the evaluation of the post-mortem decay in bones

    Journal of Proteomics 177:21–30.

    https://doi.org/10.1016/j.jprot.2018.01.016

    • PubMed
    • Google Scholar
57. 1. Procopio N
    2. Mein CA
    3. Starace S
    4. Bonicelli A
    5. Williams A

    (2021) Bone diagenesis in short timescales: insights from an exploratory proteomic analysis

    Biology 10:460.

    https://doi.org/10.3390/biology10060460

    • PubMed
    • Google Scholar
58. 1. Ramsthaler F
    2. Ebach SC
    3. Birngruber CG
    4. Verhoff MA

    (2011) Postmortem interval of skeletal remains through the detection of intraosseal hemin traces. A comparison of UV-fluorescence, luminol, hexagon-OBTI, and combur tests

    Forensic Science International 209:59–63.

    https://doi.org/10.1016/j.forsciint.2010.12.011

    • PubMed
    • Google Scholar
59. 1. Richter KK
    2. Codlin MC
    3. Seabrook M
    4. Warinner C

    (2022) A primer for zooms applications in archaeology

    PNAS 119:e2109323119.

    https://doi.org/10.1073/pnas.2109323119

    • PubMed
    • Google Scholar
60. 1. Rohart F
    2. Gautier B
    3. Singh A
    4. Lê Cao KA

    (2017) MixOmics: an R package for ’omics feature selection and multiple data integration

    PLOS Computational Biology 13:e1005752.

    https://doi.org/10.1371/journal.pcbi.1005752

    • PubMed
    • Google Scholar
61. 1. Singh A
    2. Shannon CP
    3. Gautier B
    4. Rohart F
    5. Vacher M
    6. Tebbutt SJ
    7. Lê Cao KA

    (2019) DIABLO: an integrative approach for identifying key molecular drivers from multi-omics assays

    Bioinformatics 35:3055–3062.

    https://doi.org/10.1093/bioinformatics/bty1054

    • PubMed
    • Google Scholar
62. 1. Smith PR
    2. Wilson MT

    (1990) Detection of haemoglobin in human skeletal remains by ELISA

    Journal of Archaeological Science 17:255–268.

    https://doi.org/10.1016/0305-4403(90)90023-X

    • Google Scholar
63. Book

    1. Stuart B

    (2013) Decomposition chemistry: overview, analysis, and interpretation

    In: Siegel JA, Saukko P, Houck M, editors. In Encylopedia of Forensic Sciences. AcademicPress. pp. 11–15.

    https://doi.org/10.1016/B978-0-12-382165-2.00120-3

    • Google Scholar
64. 1. Ternent T
    2. Csordas A
    3. Qi D
    4. Gómez-Baena G
    5. Beynon RJ
    6. Jones AR
    7. Hermjakob H
    8. Vizcaíno JA

    (2014) How to submit MS proteomics data to proteomexchange via the pride database

    Proteomics 14:2233–2241.

    https://doi.org/10.1002/pmic.201400120

    • PubMed
    • Google Scholar
65. 1. Wadsworth C
    2. Procopio N
    3. Anderung C
    4. Carretero JM
    5. Iriarte E
    6. Valdiosera C
    7. Elburg R
    8. Penkman K
    9. Buckley M

    (2017) Comparing ancient DNA survival and proteome content in 69 archaeological cattle tooth and bone samples from multiple european sites

    Journal of Proteomics 158:1–8.

    https://doi.org/10.1016/j.jprot.2017.01.004

    • PubMed
    • Google Scholar
66. 1. Warinner C
    2. Korzow Richter K
    3. Collins MJ

    (2022) Paleoproteomics

    Chemical Reviews 122:13401–13446.

    https://doi.org/10.1021/acs.chemrev.1c00703

    • PubMed
    • Google Scholar
67. 1. Wiley J
    2. Creamer JI
    3. Buck AM

    (2009) The assaying of haemoglobin using luminol chemiluminescence and its application to the dating of human skeletal remains

    Luminescence: The Journal of Biological and Chemical Luminescence 24:311–316.

    https://doi.org/10.1002/bio.1110

    • PubMed
    • Google Scholar
68. 1. Wilkinson DJ
    2. Rodriguez-Blanco G
    3. Dunn WB
    4. Phillips BE
    5. Williams JP
    6. Greenhaff PL
    7. Smith K
    8. Gallagher IJ
    9. Atherton PJ

    (2020) Untargeted metabolomics for uncovering biological markers of human skeletal muscle ageing

    Aging 12:12517–12533.

    https://doi.org/10.18632/aging.103513

    • PubMed
    • Google Scholar
69. 1. Wood PL
    2. Shirley NR

    (2013) Lipidomics analysis of postmortem interval: preliminary evaluation of human skeletal muscle

    Journal of Postgenomics Drug & Biomarker Development 03:1000127.

    https://doi.org/10.4172/2153-0769.1000127

    • Google Scholar
70. 1. Wu Z
    2. Lu X
    3. Chen F
    4. Dai X
    5. Ye Y
    6. Yan Y
    7. Liao L

    (2018) Estimation of early postmortem interval in rats by GC-MS-based metabolomics

    Legal Medicine 31:42–48.

    https://doi.org/10.1016/j.legalmed.2017.12.014

    • PubMed
    • Google Scholar
71. 1. Zelentsova EA
    2. Yanshole LV
    3. Melnikov AD
    4. Kudryavtsev IS
    5. Novoselov VP
    6. Tsentalovich YP

    (2020) Post-mortem changes in metabolomic profiles of human serum, aqueous humor and vitreous humor

    Metabolomics 16:80.

    https://doi.org/10.1007/s11306-020-01700-3

    • PubMed
    • Google Scholar

Author details

1. Andrea Bonicelli

   The Forensic Science Unit, Faculty of Health and Life Sciences, Northumbria University, Newcastle upon Tyne, United Kingdom

   Present address

   School of Natural Sciences, University of Central Lancashire, Preston, United Kingdom

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

   For correspondence

   ABonicelli@uclan.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9518-584X

2. Hayley L Mickleburgh

   1. Amsterdam Centre for Ancient Studies and Archaeology (ACASA) – Department of Archaeology, Faculty of Humanities, University of Amsterdam, Amsterdam, Netherlands
   2. Forensic Anthropology Center, Texas State University, San Marcos, United States

   Contribution

   Conceptualization, Supervision, Funding acquisition, Writing – original draft, Project administration, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9326-8097

3. Alberto Chighine

   Department of Medical Science and Public Health, Section of Legal Medicine, University of Cagliari, Monserrato, Italy

   Contribution

   Data curation, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0952-9712

4. Emanuela Locci

   Department of Medical Science and Public Health, Section of Legal Medicine, University of Cagliari, Monserrato, Italy

   Contribution

   Data curation, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0237-4228

5. Daniel J Wescott

   Forensic Anthropology Center, Texas State University, San Marcos, United States

   Contribution

   Conceptualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

6. Noemi Procopio

   1. The Forensic Science Unit, Faculty of Health and Life Sciences, Northumbria University, Newcastle upon Tyne, United Kingdom
   2. Forensic Anthropology Center, Texas State University, San Marcos, United States

   Present address

   School of Natural Sciences, University of Central Lancashire, Preston, United Kingdom

   Contribution

   Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Validation, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   NProcopio@uclan.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7461-7586

• 1,920

  views

• 327

  downloads

• 28

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.