Blood alternative splicing biomarker for infectious disease. Zhang et al.

Assays detecting blood transcriptome changes are being intensely studied for infectious disease diagnosis. Blood-based RNA alternative splicing events, which have not been well-characterized in pathogen infection, have potential normalization and assay platform stability advantages compared to gene expression for diagnosis. Here, we present a computational framework for developing robust alternative splicing diagnostic biomarkers. The framework analyzes patient whole-blood RNA sequencing (RNA-seq) data to identify and optimize alternative splicing biomarkers for diagnosing infectious disease. Leveraging a large prospective cohort of SARS-CoV-2 infection, we identify a major splicing program switch upon viral infection. Functional analysis reveals significant enrichment of differential splicing events in immune-specific protein domains and genes. Using test whole-blood RNA-seq data from an independent cohort, we demonstrate the superiority of alternative splicing biomarkers for SARS-CoV-2 diagnosis compared to six reported transcriptome signatures. We then optimize a subset of alternative splicing-based biomarkers to develop microfluidic PCR diagnostic assays for SARS-CoV-2 diagnosis. This assay achieves nearly perfect test accuracy (61/62=98.4%; positive percent agreement=96.7%, negative percent agreement=100%) using a naive principal component classifier, which was significantly more accurate than a gene expression PCR assay tested in the same cohort. Therefore, our RNA splicing computational framework enables a promising avenue for host response diagnosis of infection.

血液转录组变化检测分析方法（assays）在感染性疾病诊断领域正受到广泛深入的研究。在病原体感染中尚未得到充分表征的血液源性RNA可变剪接事件（RNA alternative splicing events），相较于用于诊断的基因表达分析，其在标准化流程及检测平台稳定性方面具备潜在优势。本研究提出一种用于开发稳健型可变剪接诊断生物标志物的计算框架。该框架通过分析患者全血RNA测序（RNA-seq）数据，以识别并优化用于感染性疾病诊断的可变剪接生物标志物。依托大型SARS-CoV-2感染前瞻性队列，本研究揭示了病毒感染后发生的主要剪接程序转换事件。功能富集分析显示，差异剪接事件显著富集于免疫特异性蛋白结构域及免疫相关基因中。通过使用来自独立队列的全血RNA-seq验证数据，本研究证实，相较于已报道的6种转录组特征，可变剪接生物标志物在SARS-CoV-2诊断中表现更优。随后，本研究对基于可变剪接的部分生物标志物进行优化，开发出用于SARS-CoV-2诊断的微流控PCR（microfluidic PCR）诊断检测方法。该检测方法采用朴素主成分分类器，实现了近乎完美的检测准确率（61/62=98.4%；阳性符合率为96.7%，阴性符合率为100%），其准确率显著高于同一队列中测试的基因表达PCR检测方法。综上，本研究提出的RNA剪接计算框架，为基于宿主应答的感染性疾病诊断提供了极具潜力的研究路径。