cell line: Bronchial epithelial cell line 2B4 infection: SARS-CoV-infected (MOI=0.1) time: 48 hours post infection
Extracted molecule
total RNA
Extraction protocol
total RNA extraction by using an RNAqueous-4PCR kit and following the protocol recommended by the manufacturer (Ambion, Austin, TX)
Label
Biotin
Label protocol
According to standard Affymetrix protocol.
Hybridization protocol
According to standard Affymetrix protocol.
Scan protocol
According to standard Affymetrix protocol.
Description
N/A
Data processing
For comparison across different arrays, the data for each array were normalized by Robust Multi-chip Average (RMA) using Spotfire DecisionSite (Spotfire, Inc., Somerville, MA) and GeneSifter (VizX Labs, Seattle, WA), and probe sets with expression values below the level of background noise (as determined by detection p value) were disregarded in further analyses. Results of these analyses indicated that the RNA samples used and the data generated after the normalization were of sufficiently high quality for subsequent analyses (data not shown).
For statistical purposes, all 27 arrays were analyzed as nine separate groups (mock-, SARS-CoV-, and DHOV-infected cells at 12-, 24-, and 48-hrs). The GeneSifter-based normalization was followed by pairwise comparisons of average group values and Student's t test with Benjamini and Hochberg correction. Criteria for selection of genes were as follows: All pairwise comparisons for one group versus another (e.g., mock-versus SARS-CoV-versus DHOV-infected cells at 12 hr) were expected to be at least 1.5-fold and at least 50% greater than the fold-change observed between any two controls (e.g., mock-infected replicate 1 versus mock-infected replicate 2 at 12 hr). Exceptions were made for those probe sets that were significantly altered (fold change ≥ 1.5, p value < 0.05) at an earlier or later time point, with a magnitude that exceeded the observed fold-change between replicate controls by at least 50%. This allowed for more accurate identification of those genes that increased or decreased in expression levels over time but that fell below the stringent statistical criteria at the lower transcriptional levels. In order to ensure that results obtained from comparison of SARS-CoV- and DHOV-infected cells were specific to infection, any probe set determined to be significantly different in hybridization signal based on the above criteria was further expected to have been detected as altered (fold-change ≥ 1.5 and at least 50% greater in magnitude than any change observed between control samples, p value ≤ 0.05) between mock-infected and SARS- or DHOV-infected cells. In other words, any gene with an expression level that differed between cells infected with the two different viruses was considered as irrelevant if not altered by either virus, as determined by comparison to mock-infected control cells at that same time point.