期刊名称：Proceedings of the National Academy of Sciences
印刷版ISSN：0027-8424
电子版ISSN：1091-6490
出版年度：2019
卷号：116
期号：43
页码：21350-21351
DOI：10.1073/pnas.1907705116
出版社：The National Academy of Sciences of the United States of America
摘要：Endogenous retroviruses (ERVs) are integrated retroviral elements within the human genome. Tokuyama et al. (1) recently published a computational tool, “ERVmap,” to analyze genome-wide, locus-specific expression of human ERVs. The authors found increased expression of 124 ERV loci in patients with systemic lupus erythematosus (SLE), compared to controls, and 0 down-regulated loci. In contrast, our reanalysis of their data using a Bayesian reassignment algorithm, Telescope (2), detected only 23 ERV locations with significant differential expression (DE), including 4 loci with significantly lower expression. We found that the differences between the results could be due to methodological aspects of their analysis, including alignment ambiguity, ERV annotation, and failure to account for sequencing platform as a source of variance (3). ERVmap does not adequately address the problem of alignment ambiguity, which occurs when sequencing reads align to multiple distinct genomic locations, a major challenge for quantifying expression of repetitive elements (4⇓⇓–7 … [↵][1]1To whom correspondence may be addressed.
