REPLY
In response to the comment on our recent paper (1) by Li et al., we appreciate their interest and suggestion regarding the importance of stringent bioinformatics analysis when dealing with a huge number of whole-genome probes in microarray analysis. Our analytical method essentially followed the published criteria described by Guerra et al. published in the Journal of Virology (2). Although the criteria are different from those of Li et al., we considered our bioinformatics method combined with additional distinct approaches to be appropriate for our experiments.
We have tested the robustness of our data by utilizing a different bioinformatics method reported by two other groups, Offerman et al. (3) and Cadwell et al. (4), who used the limma package (5) to identify genes that are significantly differentially expressed (P ≤ 0.05 and a ≥2-fold change). When we filtered our data sets by their methods to identify genes differentially induced by WR but not by WRΔA26 at 8 h postinfection (p.i.), the final list of 47 IFNAR-dependent secondary genes (reported in Fig. 3A and Table 3 of our paper) remained unchanged, suggesting that different statistical methods do not alter the outcome of our data interpretation. In fact, these 47 target genes were described by Tong et al. (6) as bona fide IFNAR-dependent secondary genes, giving us great confidence that the IFNAR signaling pathway is differentially induced by WR but not by WRΔA26.
Besides applying the above-mentioned statistics to enrich for differentially expressed genes, in the paper, we also utilized k-means clustering analyses that identified 36 out of 47 (76.6%) IFNAR-dependent secondary genes differentially induced by WR but not by WRΔA26 at 8 h p.i. (reported in Fig. 5 of our paper). Finally, and most importantly, the identified candidate genes and signaling pathways were experimentally verified by biological and genetic data both in vitro and in vivo. Hence, despite the different statistical methods employed, our conclusions remain valid.
Footnotes
This is a response to a letter by Li et al. (https://doi.org/10.1128/JVI.01160-17).
REFERENCES
- 1.Kasani SK, Cheng H-Y, Yeh K-H, Chang S-J, Hsu PW-C, Tung S-Y, Liang C-T, Chang W. 2017. Differential innate immune signaling in macrophages by wild-type vaccinia mature virus and a mutant virus with a deletion of the A26 protein. J Virol 91:e00767-17. doi: 10.1128/JVI.00767-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Guerra S, Najera JL, Gonzalez JM, Lopez-Fernandez LA, Climent N, Gatell JM, Gallart T, Esteban M. 2007. Distinct gene expression profiling after infection of immature human monocyte-derived dendritic cells by the attenuated poxvirus vectors MVA and NYVAC. J Virol 81:8707–8721. doi: 10.1128/JVI.00444-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Offerman K, Deffur A, Carulei O, Wilkinson R, Douglass N, Williamson AL. 2015. Six host-range restricted poxviruses from three genera induce distinct gene expression profiles in an in vivo mouse model. BMC Genomics 16:510. doi: 10.1186/s12864-015-1659-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cadwell K, Patel KK, Maloney NS, Liu TC, Ng AC, Storer CE, Head RD, Xavier R, Stappenbeck TS, Virgin HW. 2010. Virus-plus-susceptibility gene interaction determines Crohn's disease gene Atg16L1 phenotypes in intestine. Cell 141:1135–1145. doi: 10.1016/j.cell.2010.05.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Smyth GK. 2005. Limma: linear models for microarray data, p 397–420. In Gentleman R, Carey V, Dudoit S, Irizarry R, Huber W (ed), Bioinformatics and computational biology solutions using R and Bioconductor. Springer, New York, NY. [Google Scholar]
- 6.Tong AJ, Liu X, Thomas BJ, Lissner MM, Baker MR, Senagolage MD, Allred AL, Barish GD, Smale ST. 2016. A stringent systems approach uncovers gene-specific mechanisms regulating inflammation. Cell 165:165–179. doi: 10.1016/j.cell.2016.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]