血液RNAシーケンシングにおける遺伝子検査の技術的考慮事項：グロビン除去法、バッチ効果、およびサンプルタイプの評価

Blood RNA sequencing (RNA-seq) is increasingly used to enhance diagnostic yield in genetic disorders; however, technical optimization is critical for clinical implementation. We evaluated key technical considerations for blood RNA-seq, including globin depletion methods, inter-batch variability, and sample types (whole blood vs. isolated peripheral blood mononuclear cells [PBMCs]).

We compared the applicability of reagent-based globin removal (globin-RR) with bioinformatic globin removal (globin-BR) in whole-blood RNA-seq in terms of globin depletion efficiency and required sequencing depth, assessed batch effects in globin-RR samples using principal component and correlation analyses, and compared the use of whole blood versus PBMCs in terms of expression correlations and the number of detected Mendelian disease-associated genes.

Globin-RR performed better than globin-BR, achieved better coverage of clinically relevant Mendelian disease-associated genes (51.85% vs. 40.79%), and required fewer total sequencing reads (mean, ~107 vs. ~149 million) to obtain 100 million non-globin reads. Batch effects of globin-RR were low (inter-batch correlation, R²>0.96). PBMCs out-performed whole blood in terms of the number of Mendelian disease-associated genes (3,675 in PBMCs vs. 3,598 in whole blood) and percentage of highly expressed (transcripts per million >5) Mendelian disease-associated genes (54.46% vs. 51.85%) detected and therefore was the preferred sample type for RNA-seq.

Globin-RR is effective, reproducible, and practical for use in clinical whole-blood RNA-seq. With high concordance and a slightly broader gene detection range than whole blood, PBMCs are a viable alternative sample type for routine genetic diagnostic tests.