BACKGROUND Nuclear-derived cell-free DNA (cfDNA) molecules in blood plasma are nonrandomly fragmented, bearing a wealth of information related to tissues of origin. DNASE1L3 (deoxyribonuclease 1 like 3) is an important… Click to show full abstract
BACKGROUND Nuclear-derived cell-free DNA (cfDNA) molecules in blood plasma are nonrandomly fragmented, bearing a wealth of information related to tissues of origin. DNASE1L3 (deoxyribonuclease 1 like 3) is an important player in shaping the fragmentation of nuclear-derived cfDNA molecules, preferentially generating molecules with 5' "CC" dinucleotide termini (i.e., 5' CC-end motif). However, the fragment end properties of microbial cfDNA and its clinical implication remain to be explored. METHODS We performed end motif analysis on microbial cfDNA fragments in plasma samples from patients with sepsis. A sequence context-based normalization method was used to minimize the potential biases for end motif analysis. RESULTS The end motif profiles of microbial cfDNA appeared to resemble that of nuclear cfDNA (Spearman correlation coefficient: 0.82, P value <0.001). The CC-end motif was the most preferred end motif in microbial cfDNA, suggesting that DNASE1L3 might also play a role in the fragmentation of microbe-derived cfDNA in plasma. Of note, differential end motifs were present between microbial cfDNA originating from infection-causing pathogens (enriched at the CC-end) and contaminating microbial DNA potentially derived from reagents or the environment (nearly random). The use of fragment end signatures allowed differentiation between confirmed pathogens and contaminating microbes, with an area under the receiver operating characteristic curve of 0.99. The performance appeared to be superior to conventional analysis based on microbial cfDNA abundance alone. CONCLUSIONS The use of fragmentomic features could facilitate the differentiation of underlying contaminating microbes from true pathogens in sepsis. This work demonstrates the potential usefulness of microbial cfDNA fragmentomics in metagenomics analysis.
               
Click one of the above tabs to view related content.