   |
|
Members of the Toxicogenomics Research ConsortiumSupplementary InformationSupplementary Table 1Within and between laboratory median Pearson correlation coefficients of log intensities from standard array experiments (ranges of the correlation coefficients are given in the parenthesis). Laboratories 1-7 each carried out eight hybridizations to the two standard arrays; four that co-hybridized Liver RNA labeled with both Cy3 and Cy5 (LvsL), and four that cohybridized Liver RNA and Pooled RNA (LvsP). Each set was comprised of two dyeswapped samples. Pearson correlation coefficients were calculated between background-corrected log intensity values for all nucleotide sequences represented on all microarrays (Datasets A-D) (Fig. 2 of paper). * LvsL = standard liver RNA vs.standard liver RNA co-hybridization to the standard arrays; PvsP= standard Pooled RNA vs standard Pooled RNA comparisons made in silico using florescence intensity values for LvsP co-hybridization to the standard arrays. Download the PDF for Supplementary Table 1 (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_table_1.pdf) Supplementary Table 2Within and between laboratory median Pearson correlation coefficients of log ratios (LvsP) for standard array experiments using different preprocessing. Laboratories 1-7 each carried out eight hybridizations to the two standard arrays; four that co-hybridized Liver RNA labeled with both Cy3 and Cy5 (LvsL), and four that cohybridized Liver RNA and Pooled RNA (LvsP). Each set was comprised of two dyeswapped samples. To establish how the transcript level ratios between Liver and Pooled RNA (LvsP) varied depending on the method used for data normalization and background subtraction, we applied four different approaches to Dataset C. The highest median correlation of the LvsP log2 ratios between laboratories was found using Lowess Normalization without background subtraction. Thus, we applied Lowess Normalization without background adjustment to the gene expression measurements generated from all LvsP sample comparisons (Datasets B, C, D). We calculated Pearson correlation coefficients comparing the average expression ratios across laboratories for each transcript feature on each platform (Fig. 3 of paper). Download the PDF for Supplementary Table 2 (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_table_2.pdf) Supplementary Table 3Common Gene Elements Across All Platforms (Standardand Resident Arrays): Mapping to NIA NAP Clusters. A list of gene sequences that were common to all standard and resident arrays was identified using a combination of approaches including direct sequence and ID-based mapping (see Supplementary Methods). A description of the 12 array platforms is provided in the paper (Fig. 1 and Methods). Download the PDF for Supplementary Table 2 (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_tabel_3.xls) Supplementary Table 4A-FSupplementary Table 4A-F. Percent overlap of significantly induced andrepressed genes across laboratories for the Dataset D (4A, B) and Dataset C (4D, E) and number of gene transcripts identified as differentially expressed acrosslaboratories for Dataset D (4C) and Dataset C (4F). Significantly induced and repressed transcripts on all array platforms were identified. Percent overlap of significantly induced or repressed genes was calculated based on pair-wise comparisons between the laboratories participating in the development of Datasets D and C (see Supplementary Methods for description of methods used to identify significantly induced and repressed genes). Download the PDF for Supplementary Table 4A-F (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_table_4A-F.pdf) Supplementary Table 5Percentage of the functionally-enriched GO Nodes (EASE score £ 0.05) that demonstrate different levels of concordance within and between branches of the clustering dendrogram (Fig. 6 of the paper). Significantly induced and repressed transcripts on all array platforms were identified and mapped to GenBank accessions for analysis with Expression Analysis Systematic Explorer (EASE) to determine enriched Gene Ontology (GO) nodes. Greater than 50 percent of the functionally-enriched GO nodes, illustrated in Fig. 6 of the paper, have 70% concordance within Branch 2. Only 50% concordance is seen in Branch 3, and less then 50% is seen across all 3 branches. Branch 2 primarily represents Dataset D, whereas Branch 3 contains data from the Standard Spotted Arrays as well as multiple "resident" platforms. The decline in concordance seen across all datasets is likely due to the impact of Branch 1 data, which has relatively little GO node enrichment (see Supplementary Methods for description of methods used to calculate EASE scores for GO node enrichment). Download the PDF for Supplementary Table 5 (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_table_5.pdf) Supplementary Figure 1Clustering of laboratory/platform combinations based onlog ratio values associated with the common genes. Hierarchical clustering of both the laboratory/platforms and the 502 genes common to all platforms using the log10 ratio values. The relationship between the color intensity and log ratio value is illustrated by the color bar. The laboratory/platform is denoted by the letter/number combination at the bottom of every column. C = Dataset C (Standard Spotted array with data extracted from a common image analysis software package), D = Dataset D (Standard Commercial array), R = Resident array. The number details which of the 8 laboratories performed the hybridizations (see Figure 1). Resident arrays are the further described by the type of array they are cDNA = spotted cDNA, oligo = spotted oligonucleotide, C#1= commercial oligonucleotide arrays from Affymetrix, C#2 = commericial oligonucleotide arrays from Agilent, and C#3 = commercial oligonucleotide arrays from CodeLink. Download the JPEG for Supplementary Figure 1 (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/supp_fig_1.jpg) (202KB) Supplementary MethodsStandard protocols used by all laboratories for 1) Tissue Extraction and RNA Preparation, 2) Quality Control Genes,3) Production of Labeled cRNA from Total RNA, 4) Hybridization and Washing of Standard Commercial Arrays, 5) Scanning with the Agilent G2565AA Scanner for Standard Commercial Arrays, 6) Image Analysis of Standard Spotted Arrays, 7) Image Analysis of Standard Commercial Arrays, 8) Identifying Common Gene Sequences Represented on Microarray Platforms, 9) Gene Expression Data Preprocessing, 10) Statistical Analysis of Variance (ANOVA), 11) Scoring and Evaluating Gene Ontology (GO) Categories. Resident (in-house) protocols used by individual laboratories for labeling and hybridization, scanning and image analysis. Download the PDF for Supplementary Methods (http://www.niehs.nih.gov/research/atniehs/core/microarrays/trc/doc/suppmethods.pdf) Download the Primary Data
|
|
   |
|
| 