ИСТИНА

Introduction: The isolation of small non-coding sperm RNA (sncRNA) and library preparation for NGS challenging from semen with a small quantity of RNA present in sperm and the need to use DNAse. This was addressed by comparing two sncRNA library preparation kits. Patients and Methods: 24 sperm samples from the Russian Children’s Study biobank were prospectively collected at 18-20 years of age. After density gradient centrifugation sperm RNA was extracted and treated by Qiagen DNAse and by TurboDNAse. 24 pairs of NEBNext and NEXTFlex sncRNA libraries were constructed and sequenced on NextSeq 500. Unique sncRNA reads were profiled, included miRNA, piRNA and tRNA, and the 8 pairs that had >240 different sncRNAs with >=10 counts were analyzed. Variance stabilizing transformation (VST) was applied to the sncRNA counts for Bland-Altman (BA) analysis. Results: Additional treatment by Turbo DNase decreased the yield of RNA by 42%. The median input sperm count and miRNA yielded was 28.4 (15.9-44.7) mill and 0.65 (0.31-13.1) ng/mill, respectively. Although level of input miRNA measured by Qubit in both the NEBNext and NEXTFlex libraries was the same, 5.2 (2.5-105) ng, the concentration of DNA in library was significantly higher using NEBNext, 11.8 (4.9-21.3) nM, in comparison with NEXTFlex, 0.2 (0-2.1) nM. Low library quantity was critical for successful sequencing. The minimum number of unique reads for a sample to be included in that study was set at >250,000. This was observed in 11 (46%) samples prepared by NEXTFlex and 20 (83%) by NEBNext. According to BA analysis of 8 pairs that met sequencing quality, the average concordance correlation coefficient was 0.52 for miRNA, 0.58 for piRNA and 0.65 for tRNA, p<0.05. Conclusion: Using NEBNext kit we able to generate libraries with higher number of reads uniquely mapped to sncRNAs. The highest reproducibility between kits was found for tRNA and lowest for miRNA. Funding: RSF #18-15-00202; for parent RCS - NIEHS #R01 ES014370.