This study comprised of a group of four collaborators of our institute, henceforth referred to as A, B, C, and D. The participants were chosen based on their shedder status, known from previous internal studies. The group shared common work-related spaces but were neither partners nor genetically related and did not live together. All participants provided written informed consent.

Every morning, upon arrival at work, DNA from their dominant hand (right for all participants) and from the corresponding sleeve cuff of their outer-most layer of clothing (e.g., jacket or coat) were sampled. An additional DNA sample from the dominant hand was taken during the afternoon after they had spent several hours at the workplace (approximately two hours after the lunch break, resulting in a 5- to 7- h interval).

Samples were collected during 25 days, between March 6 and May 26, resulting in a total of 200 hand samples and 100 sleeve cuff samples. At this time of the year, average monthly temperatures ranged between 6.7 °C in March and 13.8 °C in May, with extreme values between − 4.2 °C and 25.6 °C. While in previous studies no significant differences were found regarding hand dominance [17,18,19], it was decided that using the dominant side was more appropriate since it is prone to be in contact with more objects, especially tools, and may therefore have the greater potential for case-relevant DNA transfer. Each participants’ daily routine was left unaltered. However, they were advised to avoid washing their hands upon arrival in the morning, before the sampling.

Participants’ activities preceding the morning sampling were recorded by a questionnaire. Questions included were mode of transport to the workplace, time since last handwashing before sampling, and whether gloves were worn more or less than 30 min before sampling in the afternoon. Lastly, the composition of the sampled sleeve cuff and the time period since the last washing of the piece of clothing were recorded. All DNA samples were collected by the same person equipped with lab coat, facemask, and gloves to minimize risk of contamination.

Buccal swabs from permanent co-residents of the participants were collected to obtain reference DNA profiles after written informed consent. A had two flatmates, B lived with their partner, C lived with their partner and their two children, and D lived alone. All reference samples received a non-speaking identifier, and they were processed anonymously in the lab. Samples were discarded immediately after profile generation. DNA profiles from the four participants and all other collaborators of the department were already available in the laboratory for quality assurance purposes.

DNA was collected using one sterile viscose forensic swab per hand (Sarstedt, Nümbrecht, Germany). Contact with external objects or surfaces at the crime scene usually occurs in the absence of liquids, which might enhance DNA transfer. Hence, dry swabs were used for sampling to mimic a more realistic scenario. DNA collection started by swabbing the palm, followed by the fingers, starting with the fifth digit and ending with the thumb.

DNA samples from sleeve cuffs were collected using SceneSafe Fast™ minitapes (SceneSafe, UK), as previously proposed by Hess and Haas [20]. No instructions regarding use and composition of the garments were imposed on the participants. Sleeve cuffs were sampled following three lines with seven applications each, resulting in an area of 13.3 cm × 6.6 cm, located on the back of the cuff (Suppl. Figure 1). DNA-Tapes were stored in 2-mL tubes at room temperature.

All laboratory analyses were conducted according to the established standard operating procedures of our lab for swabs and DNA tapes. PrepFiler Express™ Kit (ThermoFisher Scientific, Waltham, MA, USA) was used to extract DNA from swabs. Swab heads were removed with sterile razor blades and placed in 2-mL tubes with 500 μL of PrepFiler Lysis Buffer and 5 μL of Dithiothreitol 1 M (DTT). Samples were homogenized in a Precellys® 24 Touch homogenizer (Bertin instruments, Montigny-le-Bretonneux, France) for 2 × 30 s at 5900 rpm, followed by an overnight incubation at 56 °C and 400 rpm on a thermo-shaker (Labgene Scientific SA, Châtel-Saint-Denis, Switzerland). Tube contents were transferred to tubes including spin baskets for centrifugation to separate the cell lysate from the swab heads. Finally, DNA purification was performed using an AutoMate Express™ Nucleic Acid DNA Extraction System (Thermo Fisher Scientific, Waltham, MA, USA), with an elution volume of 50 μL. Extracted DNA was stored at 4 °C before subsequent analyses.

Stoop et al. [21] demonstrated that DNA recovery from the SceneSafe Fast™ Tapes by extraction with phenol–chloroform is more efficient than using magnetic bead-based extraction protocols. Therefore, tapes were cut in 10–12 pieces with sterile razor blades and incubated in a mix of 450 μL of stain extraction buffer (SEB; pH 8.0; 10 mM Tris, 10 mM EDTA, 100 mM NaCl), 50 μL of sodium dodecyl sulfate (SDS; 10%), 10 μL of proteinase K (20 mg/mL), and 10 μL of dithiothreitol (DTT; 1 M). Samples were incubated at 56 °C overnight on a thermo-shaker at 400 rpm (Labgene Scientific SA, Châtel-Saint-Denis, Switzerland). Then, 5 μL of proteinase K (20 mg/mL) was added, followed by an additional incubation of at least 2 h. The pieces of tapes were removed from the cell lysate and transferred to tubes including spin baskets for centrifugation. After adding 800 μL of phenol:chloroform:3-methylbutane-1-ol 25:24:1 (Sigma-Aldrich, US) to the lysate, the aqueous phase was transferred with distilled water in a Vivacon® 2 ETO column (Vivaproducts, Inc., US) for cleaning, as described previously [21]. Extracted DNA was stored at 4 °C before subsequent analyses.

DNA quantification was performed on a 7500 Real-Time PCR System (Thermo Fisher, Waltham, MA, USA), with the Quantifiler™ HP Kit according to manufacturer’s instructions. Results from qPCR were analyzed using the HID Real-Time PCR Analysis Software version 1.2.

Multiplex PCR was performed in reaction volumes of 25 μL using the AmpFLSTR™ NGM Select™ PCR Amplification Kit (ThermoFisher, Waltham, MA, USA), on a T3000 Biometra Thermocycler (Analytik Jena, Jena, Germany). Samples were amplified with 30 cycles by default, but 32 cycles were applied for concentrations below 0.02 ng/μL following standard internal procedures for routine casework. The optimal input amount of DNA required for multiplex PCR is 0.5 ng; therefore, a maximum sample volume of 10 μL was used for samples with concentrations below 0.05 ng/μL. For cost reasons, a single amplification was performed for all samples. For real casework samples, the Swiss law prescribes a second amplification [22].

PCR products were separated on a 3500 × L Genetic analyzer (ThermoFisher Scientific, Waltham, MA, USA), and the resulting profiles were analyzed using the software GeneMapper™ ID-X version 1.6 (ThermoFisher Scientific, Waltham, MA, USA). Profiles with at least 10 detected alleles were analyzed further. The maximum allele count method was used to determine the number of contributors. Profiles were deconvoluted and compared to the profiles from all the participants, the participants’ co-residents and from all collaborators of the department through the database search function in STRmix™ (version 2.9.1), using a Swiss reference population [23] and a FST value of 0. A real contribution to the profile (i.e., an inclusion) was assumed for all individuals with likelihood ratios (LRs) over 1000. During the deconvolution, STRmix™ estimates the respective mixture proportions of the different contributors. Using the “LR from previous” function, we compared all deconvoluted trace profiles to the profile from the respective participant from which they were sampled, to estimate the percentage of self-DNA and non-self-DNA in the samples. If STRmix proposed for a three-person mixture, e.g., mixture proportions of 65%, 25%, and 10%, and the reference profile of the participant has been attributed by the software to the 65% proportion; then, self-DNA for this sample was counted as 65% and non-self-DNA as 35%.

In forensic casework, evaluating matching profiles through probabilistic genotyping is only possible if the respective reference person, or more specifically the suspect, is known. However, this is often or even mostly not the case. Therefore, the purpose of most DNA profiles established from crime scene samples is to run a candidate search in a national DNA database. Regulated by law, Switzerland has defined the 16 STR loci, amplified by the AmpFLSTR™ NGM Select™ Kit, as database loci [22]. In the present study, we focused on major components originating from a single person, for which the criteria for an entry on the Swiss CODIS (Combined DNA Index System) database is a minimum of six typed loci. To consider loci of a single major profile as reliably typed, we used the deconvolution function of the software STRmix™ to call major components without a potential investigator bias. The probability threshold to assign an individual genotype at a given locus was kept at 99%. From internal experience, we know that a 99% probability threshold to call genotypes in the STRmix deconvolution tends to be more stringent than a manual assessment by DNA experts. Therefore, we would expect slightly more major components assigned manually by an expert in real casework, and more reliable allele calling due to PCR replicates. In addition, we did not assess our data for CODIS suitable two-person mixtures, which could also enter the Swiss national DNA database and which in real-case scenarios, could also provide us with additional investigative leads.

Statistical analyses were performed with R version 4.3.0, in Rstudio [24], using the packages “ggplot2,” “tidyverse,” “readxl,” “gdata,” “hrbrthemes,” “viridis,” and “ggpubr.” DNA amounts were log10-transformed and Shapiro–Wilk tests were performed to check for normal data distribution. Depending on the number of variables, t-tests or analyses of variance (ANOVA) were performed to check for significant differences in the detected DNA amounts. Samples that yielded an undetected amount of DNA were assigned a value of 0.0025 ng (− 2.6 on a log10 scale), corresponding to half the quantification limit of the Quantifiler™ HP Quantification Kit.