IntroductionThe Hakamata case involved the murder of four family members of a managing director who worked at a miso processing factory in Shizuoka Prefecture in 1966, along with theft and arson. A factory employee, Iwao Hakamata, was arrested and denied the charges. However, his death penalty conviction was finalized in 1980 by the Supreme Court of Japan. The first request for a retrial was filed in 1981 but ended in 2008 when the Supreme Court dismissed a special appeal. In 2008, a second request for retrial was filed with the Shizuoka District Court, which in 2014 decided to start a retrial and suspended his execution and detention. Although the Tokyo High Court overturned this decision in 2018, the Supreme Court reinstated it in 2020, leading to further hearings. In March 2023, the Tokyo High Court supported the Shizuoka District Court’s retrial decision, and a retrial began in October 2023. On September 26, 2024, the defendant was declared not guilty after judges concluded that crucial evidence and the investigation records had been fabricated, with the prosecution choosing not to appeal (https://www.japantimes.co.jp/news/2024/09/26/japan/crime-legal/hakamata-retrial-ruling/).Fourteen months after the murder, five pieces of blood-stained cotton clothes were found in a tank over a meter deep containing approximately eight tons of miso (fermented soybean paste). Testimonies and photographs taken at the time the clothes were discovered indicated that the blood stains were reddish in color. At the initial trial 45 years ago, the redness of the blood stains was not considered, and the clothing items were identified as those the defendant had been wearing at the time of the crime. As a result, the defendant was convicted. However, prior to the trial at Shizuoka District Court in 2014, the defense argued that blood stains would turn dark brown in a short period and would no longer remain red if placed in miso for over one year. Finally, in 2023, the Tokyo High Court decided to retry the case, considering it impossible for the blood stains to have remained red. The prosecutor asked one of the authors to prepare a scientific opinion on this issue and testify at the retrial, which began in 2023 (Shizuoka District Court, judgment date: September 26, 2024: https://www.courts.go.jp/app/hanrei_jp/search1).Because it is challenging to determine whether a bloodstain remains red when placed in miso for > 1 year, we considered the formation of methemoglobin, which is generated through hemoglobin oxidation, and the subsequent degradation of hemoglobin as key events in the discoloration of blood stains1,2,3. Therefore, as a first step in our evaluation, we searched the literature on dissolved oxygen (DO) concentration in miso but found no relevant information. However, DO concentrations had been measured in other Japanese fermented foods, such as shoyu (soy sauce) and sake (Japanese rice wine)4,5,6.Koji is a fermented product prepared by growing Aspergillus species in soybeans and/or grains such as rice and barley7,8. Miso is a traditional Japanese food made by mixing soybean pastes with rice koji, barley koji, soybean koji, and salt. It is then ripened for several months to over a year7,8. Aspergillus oryzae is most commonly used as a starter culture to prepare koji for making miso, and lactic acid bacteria and salt-tolerant yeasts, mainly Zygosaccharomyces rouxii, play important roles in miso maturation7,8. The koji mold in miso dies within a short time after preparation because of the high salt concentration (approximately ≥ 10%) in the mixture; however, enzymes present in koji, such as proteases, aminopeptidases, carboxypeptidases, L-glutaminases, amylases, and cellulases, remain functional over a long period7,8. These enzymes digest the starch and proteins found in grains and soybeans7,8. Although sake is fermented using rice koji made by growing A. oryzae, the type of yeast (Saccharomyces cerevisiae) used for fermentation differs from that used to make miso and shoyu9. Shoyu is produced by growing Aspergillus sojae or A. oryzae using only barley koji; however, the lactic acid bacteria and yeasts are very similar to those used in miso10. Previous studies indicated that the DO concentration in the supernatant of shoyu moromi, a soft solid formed during the fermentation process of multiple ingredients in the production of sake, shoyu, miso, and other fermented products, and a water-diluted sake moromi decreased to below the detection limit of the devices within a few days4,5,6. The detection limit was 0.1% O2 for the supernatant of shoyu moromi and 1 ppb for water-diluted sake moromi. However, in such studies, the DO concentrations in shoyu and sake were measured during brewing using the diaphragm electrode method4,5,6which is suitable for stirrable liquids but not for semisolid pastes such as miso11.One report on miso suggested that the redox potential values were affected by DO12. Other reports indicated that the redox potential of miso was close to those of shoyu and sake13,14. The redox potential of miso is the sum of those of various redox substances, and the redox state of hemoglobin in miso is affected by the type and amount of these reducing substances as well as the DO concentration in miso. This value alone does not necessarily enable prediction of the redox state of hemoglobin. Additionally, if the DO level in miso is low, oxygenation of hemoglobin is reduced, decreasing the likelihood of hemoglobin auto-oxidation and subsequent formation of methemoglobin. Taken together, the findings of these studies indicate that, although measured indirectly, the DO concentration of miso is extremely low, similar to those of shoyu and sake during brewing. Seven forensic scientists, including one of the authors, prepared an expert report on this information before the retrial began, and one of the authors testified as such at the retrial.Alternatively, the DO concentration can be measured using optical oxygen sensors, which were recently developed and mainly based on the principle of fluorescence quenching15. This method has been used to measure DO concentrations in the liquid and gas phases of wine, orange juice, and processed foods, as well as in strawberry puree and mayonnaise, which are semisolid pastes16,17,18,19,20,21,22. In the present study, we examined whether the DO concentration in miso could be measured using an optical oxygen sensor immediately after its preparation and in already-ripened miso.ResultsDO concentration in handmade MisoThe DO concentration was measured at a depth of 5 cm in each of five laboratory-made miso samples (430 g) immediately after preparation. As shown in Fig. 1, the DO concentration at the time of preparation ranged from 14.8 to 17.9% (mean ± SD = 16.8 ± 1.2%); however, these initial values were outside the measurement range of the sensor (0–5% v/v O2). Values above and below the measurement range are displayed by this sensor but are not considered accurate. The device also indicated that the O2 concentration in the air was approximately 17.2–17.9%; however, based on a previous study, this value may have been approximately 20.9%23. Furthermore, the O2 concentration (or DO concentration) in 450 mL of distilled water placed in a 500-mL beaker and left in the atmosphere was almost the same as that in air, with a DO concentration of around 8 ppm. The actual DO concentration in miso at the time of preparation was considered to be slightly higher than these indicated values. Thereafter, the DO concentration gradually decreased and fell to below the detection limit (at 0.002% O2) between 9.5 and 23.2 h after preparation (mean ± SD = 14.8 ± 4.6 h, n = 5). As this method measures the partial pressure of oxygen at the tip of the probe, the value of the oxygen partial pressure of liquids placed in the atmosphere will be the same as that in air. Thus, the detection limit (an oxygen concentration of 0.002% or less) indicates that the oxygen partial pressure in that area is 0.015 mmHg or lower. Based on this information, our optical sensor can accurately measure DO concentrations of 0.002–5% O2 in miso. The DO concentration in miso appeared to decrease to