Mitochondrial donor fraction

The proportion of donor-specific mtDNA was quantified using methods previously described²⁰. Briefly, mtDNA was extracted from pretransplant whole blood samples, amplified, underwent library preparation, and was sequenced. Processing was separate for donor and recipient samples. Raw sequencing data was trimmed and aligned to the human genome, and mitochondrial sequences were selected. We estimated the presence of each nucleotide at each position using bam-readcount (https://github.com/genome/bam-readcount), and a mitochondrial consensus sequence was determined for the donor and recipient. The consensus sequences were compared and single-nucleotide polymorphisms (SNPs) discriminating the two individuals were identified. Downstream cfDNA sequence data aligning to the mitochondrial were compared to the bases with SNPs that discriminate the donor and recipient. For each SNP position across the mitochondrial genome, we determined the donor fraction by dividing the donor SNP count by the total number of donor and recipient counts at the SNP. We discarded a donor fraction estimation at a point if the depth of sequencing was <50×. We determined the mean of the estimated donor fraction at all SNPs to quantify the mitochondrial donor fraction. One sample was removed owing to low depth of sequencing across all SNPs.