In the March 22, 2023, Current Biology publication on analyses of Beethoven’s genome, “Genomic Analyses of Hair from Ludwig van Beethoven,” we revealed that the highly publicized Hiller Lock, analyses of which had earlier led to conclusions that (1) Beethoven suffered from plumbism and (2) was administered neither mercury nor opiates, had in fact originated from a woman.
As the Hiller Lock, also known as the “Guevara Lock,” is definitely inauthentic, we therefore stated that the conclusions drawn from analyses of the Hiller Lock no longer apply to Beethoven:
“Toxicological analyses of hairs extracted from this lock [the Hiller Lock] have been used to argue that Beethoven’s health problems were caused or compounded by plumbism, and to refute suggestions that he was administered opiates during the course of his final illness and mercury for a hypothesized infection with syphilis. We now conclude that these findings no longer apply to Beethoven.”
(Begg et al., 2023, p. 3)
Later in our paper, while discussing possible causes of Beethoven’s hearing loss, we stated that substantiation of the lead poisoning theory would have to await toxicological analyses of authenticated samples:
“Substantiation of the previously hypothesized role of plumbism as a causative or contributory factor to Beethoven’s hearing loss must await analyses of samples authenticated via genetic testing.”
(Begg et al., 2023, p. 11)
We exercised care to avoid dismissing theories concerning the presence or absence of lead, mercury, or opiates in Beethoven’s hair, instead endorsing future toxicological analyses using authenticated samples. When our study was published, there were no toxicological analyses of hair samples that had been authenticated through genomic testing, and thus no dependable empirical data to either support or refute these theories.
However, we did refute one conclusion stemming from an earlier analysis of the Hiller Lock. In 2007 Professor Christian Reiter claimed to have observed congruent longitudinal distributions of heavy metal isotopes shared between the Hiller, Halm-Epstein, and Erdödy Locks, claiming to have thereby proven their authenticity. We stated that this conclusion must have been in error, as the Hiller Lock is inauthentic:
“We additionally demonstrate that patterns of longitudinally distributed lead isotope concentrations believed to have been shared between hair strands from the Hiller, Halm-Epstein, and Erdödy Locks do not constitute proof of their authenticity, as the Hiller Lock is inauthentic.”
(Begg et al., 2023, p. 3)
Here is an English translation of the statement we refuted:
“Recently a third lock from the estate of the composer Anton Halm (1789-1872) that Halm had bequeathed to his pupil Julius Epstein (1832-1926) was investigated with regard to its authenticity. The pattern of its heavy metal distribution was found to be congruent with the pattern characteristic of Beethoven’s discovered in the Guevara and Erdödy locks. Thus the analytical results presented here not only provide new clues regarding the circumstances and causes of Beethoven’s death but prove to be a reliable method for identifying doubtful locks of Beethoven’s hair.”
(Reiter, 2007, p. 5) [translated by Michael Lorenz]
In response to these findings, Professor Reiter has made repeated suggestions to journalists that the DNA we sequenced from the Hiller Lock was contamination, intimating that the Hiller Lock hairs were nonetheless authentic, and his conclusion still valid. This was despite the fact that the results of contamination testing of the Hiller Lock were published in our study, yielding the lowest possible final mitochondrial contamination estimate, using the algorithm ‘Schmutzi,’ of 0-2%. While some journalists declined to include Professor Reiter’s suggestion of contamination in their reporting, unfortunately, several highly influential media outlets failed to fact-check Professor Reiter’s suggestion by querying our published contamination testing result for the Hiller Lock, and widely reported Professor Reiter’s unsupported suggestion.
I was recently contacted by a journalist, once more asking me how I can be so sure that the DNA we sequenced from the Hiller Lock is not contamination. I unhappily learned that Professor Reiter is still suggesting that our Hiller Lock data was contaminated.
Regrettably, I feel that I am obliged to publicly respond to Professor Reiter’s suggestion of contamination in the Hiller Lock dataset.
Professor Reiter proposes that the locket in which the Hiller Lock hairs were stored had previously contained hairs from a woman (according to the journalist I spoke to, he posits Ferdinand Hiller’s mother). These hairs were supposedly replaced with authentic Beethoven hairs. Trace levels of DNA contamination from this earlier, hypothetical lock purportedly remained on the interior surface of the locket, and were then transferred onto the surface of authentic Beethoven hairs, which both Professor Reiter and I independently tested. These trace quantities of contaminant DNA on the surface of the tested hairs apparently survived decontamination in a state-of-the-art ancient DNA cleanroom, coming to completely overwhelm any evidence of authentic Beethoven DNA in our sequence data for the Hiller Lock.
Professor Reiter’s suggestion that an earlier lock of hair from a woman was once held within the same locket is speculation, lacking any historical support. It should also be noted that the discovery within this locket of fragments of an earlier inscription, penned prior to the locket’s refurbishment in 1911, and bearing the words “Ferdinand Hiller,” “Beethoven,” and “abgeschnitten,” might be more consistent with the converse of Reiter’s theory: That authentic Beethoven hairs might have been replaced with inauthentic hairs from a woman, and not vice versa.
Regardless, we are primarily concerned with the DNA sequence data from which we drew our conclusions in our 2023 publication. Below, I am going to present the reader with a close examination of the mitochondrial DNA sequence data from the Hiller Lock.
It is first necessary to state that the Hiller Lock sequence data, and all of the sequence data analyzed in our 2023 study, were generated by pairing state-of-the-art ancient DNA laboratory and bioinformatics methods with library-based Illumina next-generation sequencing technologies. This combined approach represents a significant departure from more conventional forensic genetics techniques. The reason for this departure is that hair shafts as a DNA source are not well suited to conventional forensic genetics methods. During the formation of hair shafts, the DNA which ends up being incorporated into the hair shaft is enzymatically broken down into extremely short fragments. Consequently, more commonly applied forensic genetics methods, which usually amplify large and specific DNA molecules, followed by low-efficiency (or “low-throughput”) sequencing methods, typically fail to yield positive results when applied to hair shafts. Our approach, however, was tailor-made for the efficient recovery of large quantities of the highly-fragmented, “ultra-short” DNA molecules characteristic of both modern and historical hair shafts. Our approach further differs from more conventional forensic genetic approaches because, instead of attempting to target and sequence large and specific DNA molecules (e.g. using “Polymerase Chain Reaction” followed by Sanger Sequencing), we indiscriminately sample all of the DNA extracted from the hair shafts (colloquially termed “shotgun sequencing”), and then algorithmically align these DNA molecules to a standardized reference genome. After a sufficient number of unique DNA molecules have been very confidently aligned to a particular position on this reference genome, we then have a sufficiently redundant number of independent observations at each genomic position (a.k.a. “depth of coverage”) to be able to determine the genotype (e.g. a mutation) at a given position. Where the amount of data is sufficient, this dual-approach offers the added potential of accurately quantifying levels of DNA contamination. Unfortunately, an out-dated adage that hair shafts are not viable as a DNA source is still quite widely held. This adage was simply never correct. As we demonstrated in our 2023 study, using the appropriate techniques, sufficient quantities of DNA can be recovered from small quantities of historical hair shafts for the purpose of authentication testing, even permitting the sequencing of high-coverage genomes.
Of the eight locks of hair attributed to Beethoven that we tested, the Hiller Lock was the best-preserved sample in terms of the quantity of DNA that was preserved in the hairs, and the relatively low levels of DNA damage. This sample was furthermore subjected to a very large amount of sequencing. Altogether, we obtained a volume of data far in excess of what was required for authentication analyses.
The upshot of this large amount of sequence data is that the mitochondrial genome was reconstructed in its near-entirety many hundreds of times over. A mitochondrial genome is “haploid,” meaning that at any given position along its length, all of the sequenced DNA molecules aligning to a given position should show the same nucleotide. The great depth of coverage for our Hiller Lock mitochondrial genome, paired with its haploid nature, means that it should theoretically permit the quantification of even the most minute traces of contamination from a second individual.
Given that Professor Reiter is suggesting that the Hiller Lock hairs are from Beethoven, and that his theory would logically hold that both the Beethoven hairs, and any older trace contamination, would have shared identical storage conditions inside the locket for the duration of their shared history, it should also logically hold that we should expect to see some evidence of Beethoven’s DNA within this considerable dataset in support of his theory.
Beethoven’s mitochondrial haplogroup is H1b1+16,362C, distributed broadly throughout Western and Central Europe, and the Hiller Lock’s mitochondrial haplogroup is K1a1b1a, a major Ashkenazi Jewish sub-clade. These are very different haplogroups, estimated to have last shared a common matrilineal ancestor approximately 70,000 years ago. We therefore have a fair number of positions along Beethoven’s mitochondrial genome that differ distinctly from those in the Hiller Lock. So, are there any DNA molecules definitely containing mutations consistent with Beethoven’s mitochondrial haplogroup within the Hiller Lock dataset?
To answer this, I am going to examine the Hiller Lock mitochondrial DNA sequence data aligning to each of the positions harboring Beethoven’s clade-defining mutations. In each case, Beethoven and the Hiller Lock have distinctly different mutations. If even one of the DNA molecules in the Hiller Lock dataset aligning to one of these positions bears a mutation consistent with Beethoven’s haplogroup, there might arguably be some support for Professor Reiter’s suggestion.
In order to afford the least possible ambiguity in this comparison, I impose several quality filtering measures. I ask the reader to forgive me for the following technical details: I first impose a stringent mapping quality filter for these DNA fragments of MAPQ ≥ 30. I then restrict this analysis to positions where each clade-defining mutation for Beethoven is, relative to the consensus base call at the same position in the Hiller Lock dataset, neither a “cytosine to thymine” (C to T) transition, a “guanine to adenine” (G to A) transition, nor at a known poly-C stretch or indel (insertion/deletion) site, as transitions at these positions which might have arisen from contamination will be indistinguishable from C->T or G->A transitions arising from hydrolytic deamination (“DNA damage”) and its reverse-complement, or from indels or poly-C stretches. We nonetheless retain a list of the following six positions for investigation:
11719 (A to G)
14766 (T to C)
7028 (T to C)
16356 (T to C)
3796 (A to G)
16362 (T to C)
While excluding the following five positions:
73 (G to A)
2706 (G to A)
3010 (G to A)
16189 (indel)
16176 (C to T)
The results of this analysis of the Hiller Lock sequence data at these six positions are tabulated below:
There is not one unambiguous molecule of Beethoven’s DNA in the entire quality-filtered Hiller Lock mitochondrial dataset. This even includes positions with quite extraordinary depth of coverage, such as those at positions 16,356 and 16,362, where not a single one of the over 800 unique DNA molecules aligning with high mapping quality to either position contains a clade-defining mutation leading to Beethoven’s haplogroup of H1b1+16,362C.
In order for all of our Hiller Lock data to be contamination, this older, residual, trace contamination, sharing identical storage conditions with the Hiller Lock subsequent to the purported swapping of this hypothetical lock with Beethoven’s hair, and subsequent to surviving decontamination, would then need to outnumber any of Beethoven’s actual DNA by almost three orders of magnitude.
No historical hair sample I analyzed in the course of either my Master’s thesis, which focused on ancient and historical hair as a DNA source, or in my doctoral dissertation research, ever exceeded a final mitochondrial contamination estimate of 0-3% after decontamination. This was irrespective of whether we were analyzing bulbs or rootless hair shafts, the amount of hair used per extraction, or whether the hairs were decontaminated in a low-concentration (0.5%) bleach dilution, or multiple immersions in sterile water. In order to reconcile Professor Reiter’s theory with these findings, the ratio of contamination he is alleging would have to be of a truly unprecedented scale. The Hiller Lock would need to be the most singularly bizarre and extreme outlier of any sample that I have ever analyzed, and in all contravention of the contingencies that are logically premised by Professor Reiter’s theory.
The absence of any ambiguity here is reflected in all of the other observations of the Hiller Lock dataset. The coverage on the sex chromosomes is consistent with the Hiller Lock having two X-chromosomes, but no Y-chromosome, demonstrating that the sample derives from a woman. Analyses of autosomal relatedness between all tested samples, which demonstrate that five of our samples originate from a single individual (Beethoven), reveal no close degree of relatedness between any of the Beethoven samples and the Hiller Lock. And, analyses of autosomal ancestry show that, while Beethoven groups among individuals with Western or Central European ancestry, the Hiller Lock lands among individuals with North African and Levantine ancestry, including several Jewish diasporas, ostensibly consistent with the Ashkenazi Jewish mitochondrial haplogroup assignment.
It should finally be noted that the newly published results of toxicological examinations of authenticated Beethoven hair samples, which includes the Halm-Thayer Lock (which perhaps was split from the same original as the Halm-Epstein Lock, which Reiter had earlier compared to the Hiller Lock), are not “congruent” with the conclusions drawn from earlier testing for heavy metals in the Hiller Lock. While an earlier analysis by William Walsh’s laboratory did not detect any traces of mercury in the Hiller Lock, leading to the conclusion that Beethoven had not been administered mercury-containing medicines, traces of mercury are detected in the authenticated Halm-Thayer and Bermann Locks.
It is my hope that sufficient reassurance has been provided to any concerned or confused reader that the Hiller Lock does not exhibit even one molecule of contamination with any DNA that might have originated from Beethoven, and does not originate from Beethoven. Professor Reiter’s repeated suggestion that the Hiller Lock data we analyzed is contamination originating from an earlier lock of hair is an unfounded speculation, premised upon an ahistorical revision of that lock’s provenance history, and which totally contravenes all of the robustly supported and published observations of the genetic data generated from that sample.
As Professor Reiter did not provide any empirical support for his claimed observation of congruent longitudinal heavy metal distributions shared between the Hiller Lock, and the Halm-Epstein and Erdödy Locks, I do not know how he arrived at that conclusion, but, to reiterate the statement in our 2023 publication, Professor Reiter’s 2007 conclusion is in error.
Bibliography
Tristan J.A. Begg et al., “Genomic Analyses of Hair from Ludwig van Beethoven,” Current Biology, March 22, 2023, p. S0960982223001811. Available at:
https://doi.org/10.1016/j.cub.2023.02.041
Christian Reiter, “The Causes of Beethoven’s Death and His Locks of Hair: A Forensic-Toxicological Investigation,” The Beethoven Journal 22, no. 1(2007): 2–5.