Quantasoft

Plasma cfDNA and cffDNA quantification was done by a ddPCR assay designed to amplify a short segment (140 bp) of human RASSF1A gene promoter region. Forward primer 5′- AGT GCG CGC GTG AGT AGT -3′ and reverse primer 5′- GGC GAA AGT AAC GGA CCT AGT-3′ were designed using Roche ProbeFinder online software. Probe for this assay is Roche’s universal probe library probe number 24 (cat. no. 04686985001) which was recommended by the ProbeFinder. Primers were purchased from Integrated DNA Technologies (IDT) (Coralville, IA). Universal probe number 24 was purchased from Roche. A PCR master mix, 2× ddPCR™ Supermix for Probes, was purchased from Bio-Rad Laboratories (Hercules, CA). Final concentrations of primers and probe in PCR reactions were 900 nM and 250 nM, respectively, in a final volume of 20 μL. The DNA template input volume was 5 μL. A Bio-Rad QX200 Droplet Digital™ PCR System was used as described by Hindson and colleagues [19 (link)]. Thermal cycling was performed with a Bio-Rad C1000 Touch Thermal cycler. The following PCR conditions were used: 95°C for 10 min, 40 cycles of 30 s at 94°C and 50 s at 60°C followed by a heating step at 98°C for 10 min to inactivate the polymerase. Data analysis was performed using Bio-Rad QuantaSoft software version 1.7.4.0917.

Plasma from the healthy donors was used for establishing the limit of blank. They were split into a discovery cohort and a validation cohort. Plasma from the 50 donors in the discovery cohort was analyzed for meth-HOXA9. Results were exported from the QX200 Droplet Digital Reader (Bio-Rad, Hercules, CA, USA) as the number of droplets containing meth-HOXA9 detected by ddPCR, and a <5% false-positive rate was arbitrarily decided as realistic. This resulted in a cut-point of ≥5 droplets containing meth-HOXA9 equaling a positive test, as this was true of one healthy donor (2%). The validation cohort of the 50 other donors showed exactly the same results of one donor with ≥5 droplets containing meth-HOXA9. Hence, the limit of blank was set at 4 droplets, and the limit of detection was set at 5 droplets.
After setting the limit of blank, data were normalized to the level of the albumin gene. Meth-HOXA9 copies divided by albumin copies resulted in a fraction of meth-HOXA9. These data were exported from QuantaSoft™ (Bio-Rad, Hercules, CA, USA) as the percentage of meth-HOXA9, including a 95% confidence interval (CI) derived from a Poisson distribution [35 ].

A 20µL ddPCR master mix was prepared with QX200™ ddPCR™ EvaGreen^® SuperMix (BioRad) following the manufacturer’s instructions (BioRad), with a final primer concentration of 120nM and with 10ng of nucleic acid template. PCR was performed on Bio-Rad C1000 Touch Thermal Cycler with the following conditions: 95 °C for 5 min, 40 cycles at 95 °C for 15 s and 60 °C for 1 min, 4 °C for 5 min, 90 °C for 5 min, and incubation at 10 °C. Final products were transferred to QX200™ Droplet Reader and quantified as gene copies (per 20µL) using Bio-Rad QuantaSoft (v.1.7.4.0917). Lactobacillus and Pelomonas loads were quantified using genus specific primers on 242 DNA and cDNA samples each (Table 3).
Table 3

List of primers used for ddPCR.

	Primer	5’ – 3’	Reference
Total bacteria	63 F	GCAGGCCTAACACATGCAAGTC	[56 ]
	355R	CTGCTGCCTCCCGTAGGAGT
Lactobacillus	F-Lacto	GAGGCAGCAGTAGGGAATCTTC	[57 (link)]
	R-Lacto	GGCCAGTTACTACCTCTATCCTTCTTC
Pelomonas	357 F	CGGGTTGTAAACCGCTTTTGT
	550R	CGGGGATTTCACCTCTGTCT

The ddPCR data was analyzed using QuantaSoft version 1.7.4.0917 (Bio-Rad). The same fluorescence thresholds were applied to all samples across all plates. Wells with fewer than 10,000 droplets were removed. Concentrations of IFI6, MX1 and ISG15 were divided by the concentration of UBC from the corresponding sample to yield copies of each gene per copy of UBC. This value was log2-transformed to convert it to a normal distribution and place it on a comparable scale to the microarray data. Replicate wells were then averaged. Fold changes were calculated by subtracting the expression level from the off-treatment sample from the on-treatment sample.

Digital PCR experiments were carried out using the QX200™ Droplet Digital™ PCR System (Bio-Rad Laboratories) following the protocols described earlier [6 (link), 16 (link), 17 ]. A methodical overview of all experimental setups is presented in Supplementary Data 1. Context sequences, PCR annealing temperatures and supplier information for all assays used are provided in Supplementary Table 1.
Raw digital PCR results were acquired using QuantaSoft (version 1.7.4, Bio-Rad Laboratories) and imported in the online digital PCR management and analysis application Roodcom WebAnalysis (version 1.9.4, available via https://webanalysis.roodcom.nl).

ddPCR assay was performed on the same sample twice using the QX200^™ Droplet Digital^™ PCR System (Bio-Rad Laboratories, Hercules, CA, USA) and PCR data were quantified using QuantaSoft^™ software (Bio-Rad Laboratories) and results are expressed as fractional abundance (mutant allele frequency: MAF) for each tumor tissue sample and as copies/μL of mutant DNA for each plasma sample as described previously [22 (link)]. Our ddPCR analysis of four representative ESR1 LBD mutations (Y537S, Y537N, Y537C, and D538G) had already been optimized by comparative analysis of a dilution series of each synthetic ESR1 LBD mutant oligonucleotide as reported previously [20 ]. All samples were compared with the ESR1 WT molecule and each ESR1 mutant molecule as positive control. A water-only (no template) control and WT normal human DNA (TaqMan Control Genomic DNA) were run in parallel for each ddPCR reaction as negative control.