< published in Epigenetics and Chromatin 2015, 8:22 doi:10.1186/s13072-015-0014-8.
### 3. USING WANDERER ########################################################################################
3.1. **Enter** a gene name (BRCA1 is used as default) or the equivalent Ensembl gene ID and press the **Refresh** button to update the output. Modifying the rest of options will automatically update the graphs.
3.2. Choose a **Dataset** (default dataset is Breast Invasive Carcinoma (BRCA)) and the **Data type** (450k Methylation array or Illumina HiSeq RNAseq) in the respective drop down menus.
3.3. Use the **Zoom** sliding bar to adjust the displayed chromosomal region or tick the **Specify a region** box for a more precise selection. Only values in the range defined by the zoom slider will be accepted.
3.4. Customize the graphical display of the data using **Plotting parameters**. Graphs are automatically updated after any change.
3.5. **Download** graphs and tables containing the selected data or share the output by generating a permanent link.
### 4. DESCRIPTION OF THE DOWNLODABLE FILES GENERATED BY WANDERER ############################################
> **IMPORTANT WARNING ON USING EXCEL AND OTHER SPREADSHEET SOFTWARE TO IMPORT GENOMIC DATA**
> It is known that Excel and other spreadsheet softwares may change gene names when importing text data [(Zeeberg et al., 2004)]. A default feature will misidentify specific gene names (e.g.: SET1, AGO3) as dates. This conversion may be avoided by formatting the spreadsheet cells as Text before importing.
Wanderer generates several images and data files that may be downloaded as a compressed file. The files are named using the following nomenclature:
`Wanderer_[gene]_[data type]_[dataset]_[date and time of the query].`
#### 4.1 DNA METHYLATION FILES ###############################################################################
The predefined query for DNA methylation produces the following graphs and data files:
4.1.1. Profile plots displaying the beta values for each probe in two separate panels representing subset of normal (upper panel, blue marks) and tumor samples (bottom panel, red marks) of the selected dataset. Lines link different probes corresponding to the same sample. The gene location and direction is depicted with an arrow. CpG island located probes are colored in green.
_File name examples_:
_Raster version:_ `1_Wanderer_BRCA1_methylation_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `2_Wanderer_BRCA1_methylation_brca_Apr_29_2015_at_170440_CEST.pdf`
4.1.2. Two tables consisting of a comma separated data matrix containing data for normal and tumor samples respectively. Table contents: First column: 450K Methylation array probe ID, Rest of the columns: DNA methylation beta value for each sample.
_File name examples:_
`3_Wanderer_BRCA1_methylation_brca_Normal_Apr_29_2015_at_170440_CEST.csv`
`4_Wanderer_BRCA1_methylation_brca_Tumor_Apr_29_2015_at_170440_CEST.csv`
4.1.3. Profile plots displaying the average methylation for all the normal (blue) and tumor samples (red) of the selected dataset. The gene location and strand is depicted with an arrow. CpG island located probes are colored in green. The CpGs showing statistical differences between normal and tumor are highlighted with an asterisk (Wilcoxon adjusted p-value < adjusted pval threshold selected).
_File name examples:_
_Raster version:_ `5_Wanderer_BRCA1_Mean_methylation_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `6_Wanderer_BRCA1_Mean_methylation_brca_Apr_29_2015_at_170440_CEST.pdf`
4.1.4. A comma separated data matrix with the annotation for each of the informative probes and descriptive analysis of the DNA methylation data. The columns correspond to:
- probe, probe name
- chr, the chromosome
- cg_start, the genomic position the CpG starts at
- cg_end, the genomic position the CpG ends at
- percentgc, the GC content of the illumina 450k array probe
- probetype, the type of the illumina 450k array probe
- probestart, the genomic position the probe starts at
- probeend, the genomic position the probe ends at
- genestart, the genomic position the closest gene to the probe starts at
- geneend, the genomic position the closest gene to the probe ends at
- genestrand, the closest gene strand
- ENSEMBL_geneID, the closest gene id at ensembl
- genebiotype, the closest gene biotype (protein coding, retained intron...)
- genename, the closest gene symbol
- cpgistart, for those CpG are inside a CpG island, the coordinate this islands starts at
- cpgiend, for those CpG are inside a CpG island, the coordinate this islands ends at
- cpgiid, for those CpG are inside a CpG island, CpG island identifier
- Norm_nsamples, number of normal samples for this dataset in this release
- Norm_mean, mean of the beta values for normals
- Norm_sd, standard deviation of the beta values for normals
- Tum_nsamples, number of tumor samples for this dataset in this release
- Tum_mean, mean of the beta values for tumors
- Tum_sd, standard deviation of the beta values for tumors
- wilcox_stat, if there are enough samples, Wilcoxon Rank Sum Test W parameter (nonparametric comparison of normals vs tumors)
- pval, if there are enough samples, Wilcoxon Rank Sum Test p value (nonparametric comparison of normals vs tumors, low values indicates that differences are detected)
- adj.pval, if there are enough samples, Benjamini and Hochberg adjustement (False Discovery Rate) for multiple testing.
__File name examples:__
`7_Wanderer_BRCA1_methylation_brca_annotations_and_statistical_analysis_Apr_29_2015_at_170440_CEST.csv`
4.1.5. Plots showing the correlation between DNA methylation beta values and RNAseq expression summarized by gene for all the normal (blue) and tumor samples (red) of the selected dataset. The expression values are log2-transformed normalized RSEM values [(Guo et al., 2013)] and reflect the expression of the gene as a whole. Three methods to obtain the correlation can be choosed: Spearman's rho, Kendall's tau and Pearson's r.
_File name examples:_
_Raster version:_ `8_Wanderer_BRCA1_correl_RNAseqGeneVSmethylation_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `9_Wanderer_BRCA1_correl_RNAseqGeneVSmethylation_brca_Apr_29_2015_at_170440_CEST.pdf`
4.1.6. A comma separated data matrix with the available normal and tumors sample names (first column) and the log2-transformed RSEM values for your gene of interest (second column).
_File name examples:_
`10_Wanderer_BRCA1_methylation_brca_Normal_RNAseqGENE_CommonWithMethylation_Apr_29_2015_at_170440_CEST.csv`
`11_Wanderer_BRCA1_methylation_brca_Tumor_RNAseqGENE_CommonWithMethylation_Apr_29_2015_at_170440_CEST.csv`
#### 4.2 GENE EXPRESSION FILES ###############################################################################
4.2.1. Profile plots displaying the log2-transformed RPKM values [(Guo et al., 2013)] for each exon in two separate panels for a subset of normal (upper panel, blue marks) and tumor samples (bottom panel, red marks) from the selected dataset. Lines link values corresponding to the same sample. The gene location and strand is depicted with an arrow.
_File name examples:_
_Raster version:_ `1_Wanderer_BRCA1_expression_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `2_Wanderer_BRCA1_expression_brca_Apr_29_2015_at_170440_CEST.pdf`
4.2.2. Two tables consisting of a comma separated data matrix containing data for normal and tumor samples respectively. Table contents: First column: RNAseq exon ID, Rest of the columns:log2-transformed RPKM values for each sample.
_File name examples:_
`3_Wanderer_BRCA1_expression_brca_Normal_Apr_29_2015_at_170440_CEST.csv`
`4_Wanderer_BRCA1_expression_brca_Tumor_Apr_29_2015_at_170440_CEST.csv`
4.2.3. Profile plots of the average expression for all the normal (blue) and tumor samples (red) of the selected dataset. The gene location and strand is depicted with an arrow. The exons showing statistically significant differences between normal and tumor are highlighted with an asterisk (Wilcoxon adjusted p-value < adjusted pval threshold selected).
_File name examples:_
_Raster version:_ `5_Wanderer_BRCA1_Mean_expression_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `6_Wanderer_BRCA1_Mean_expression_brca_Apr_29_2015_at_170440_CEST.pdf`
4.2.4. A comma separated data matrix with the annotation of your gene expression data, as well some descriptive analysis. The columns correspond to:
- exon, exon identifier according to the TCGA pipeline
- id, exon identifier according to Genome Browser exons track
- ENSEMBL_geneID, the gene identifier according to ENSEMBL (ENSG identifier)
- ENSEMBL_transcriptID, the transcript identifier according to ENSEMBL (ENST identifier)
- chr, the chromosome the exon is located at
- exon_start, the genomic coordinate the exon starts at
- exon_end, the genomic coordinate the exon ends at
- strand, the genomic strand of the exon's gene.
- genestart, the genomic start position of the exon's gene.
- geneend, the genomic end position of the exon's gene.
- genebiotype, exon's gene biotype (protein coding, retained intron...)
- genename, exon's gene symbol
- rnaseqgeneid, the TCGA exon's gene identifier
- Norm_nsamples, number of normal samples for this dataset in this release
- Norm_mean, mean of log2-transformed RPKM for normals
- Norm_sd, standard deviation of log2-transformed RPKM in normals
- Tum_nsamples, number of tumor samples for this dataset in this release
- Tum_mean, mean of log2-transformed RPKM for tumors
- Tum_sd, standard deviation of log2-transformed RPKM in tumors
- wilcox_stat, if there are enough samples, Wilcoxon Rank Sum Test W parameter (nonparametric comparison of normals vs tumors)
- pval, if there are enough samples, Wilcoxon Rank Sum Test p value (nonparametric comparison of normals vs tumors, low values indicates that differences are detected)
- adj.pval, if there are enough samples, Benjamini and Hochberg adjustement (False Discovery Rate) for multiple testing.
_File name example:_
`7_Wanderer_BRCA1_expression_brca_annotations_and_statistical_analysis_Apr_29_2015_at_170440_CEST.csv`
4.2.5. Boxplot and a stripchart graphs showing the expression values summarized by gene for all the normal (blue) and tumor samples (red) of the selected dataset. The expression values are log2-transformed normalized RSEM values [(Guo et al., 2013)] and reflect the expression of the gene as a whole.
_File name examples:_
_Raster version:_ `8_Wanderer_BRCA1_boxplot_expression_brca_Apr_29_2015_at_170440_CEST.png`
_Vectorial format:_ `9_Wanderer_BRCA1_boxplot_expression_brca_Apr_29_2015_at_170440_CEST.pdf`
4.2.6. A comma separated data matrix with the available normal and tumors sample names (first column) and the log2-transformed RSEM values for your gene of interest (second column).
_File name examples:_
`10_Wanderer_BRCA1_expression_brca_Normal_RNAseqGENE_Apr_29_2015_at_170440_CEST.csv`
`11_Wanderer_BRCA1_expression_brca_Tumor_RNAseqGENE_Apr_29_2015_at_170440_CEST.csv`
#### 4.3 CLINICAL DATA #######################################################################################
TCGA clinical data is provided as **biotab** files. Biotab files are an amenable, spreedsheet-friendly data sources that reflect clinical and biospecimen information for a set of patients. In this file, each column is a clinical element and each row, a TCGA participant.
We note that biotab files cover data unique to each TCGA participant (for instance, the patient's weight and height) and therefore are independent on whether methylation or expression was scrutinized. We offer the biotab files for all the participants available for a dataset (for instance, colon adenocarcinoma), regardless of whether they were explored for expression or methylation, nor in solid primary tumors or in adjacent normal.
_File name example:_
`BRCA_Clinical__nationwidechildrens.org_clinical_patient_brca.txt`
### 5. DATA AVAILABILITY ####################################################################################
Wanderer uses a TCGA snapshot taken on July 2014. Data availability for normals (N) and tumors (T) is detailed below.
| Description | Name | Meth_N | Meth_T | RNAseq_N | RNAseq_T |
|----------------------- |:------:|:-----:|:-----:|:-----:|:-----:|
| Adrenocortical carcinoma | ACC | 0 | 80 | 0 | 79 |
| Bladder Urothelial Carcinoma | BLCA | 21 | 358 | 19 | 267 |
| Brain Lower Grade Glioma | LGG | 0 | 511 | 0 | 513 |
| Breast invasive carcinoma | BRCA | 98 | 743 | 113 | 1052 |
| Cervical squamous cell carcinoma and endocervical adenocarcinoma | CESC | 3 | 256 | 3 | 207 |
| Colon adenocarcinoma | COAD| 38 | 302 | 41 | 262 |
| Esophageal carcinoma | ESCA | 16 | 185 | 0 | 0 |
| Glioblastoma multiforme | GBM | 2 | 129 | 5 | 156 |
| Head and Neck squamous cell carcinoma | HNSC | 50 | 528 | 43 | 497 |
| Kidney Chromophobe | KICH | 0 | 66 | 25 | 66 |
| Kidney renal clear cell carcinoma | KIRC | 160 | 324 | 72 | 518 |
| Kidney renal papillary cell carcinoma | KIRP | 45 | 226 | 30 | 198 |
| Liver hepatocellular carcinoma | LIHC | 50 | 256 | 50 | 212 |
| Lung adenocarcinoma | LUAD | 32 | 463 | 58 | 488 |
| Lung squamous cell carcinoma | LUSC | 43 | 361 | 50 | 491 |
| Lymphoid Neoplasm Diffuse Large B+AC0-cell Lymphoma | DLBC | 0 | 48 | 0 | 28 |
| Mesothelioma | MESO | 0 | 37 | 0 | 36 |
| Ovarian serous cystadenocarcinoma | OV | 0 | 10 | 0 | 262 |
| Pancreatic adenocarcinoma | PAAD | 10 | 146 | 4 | 96 |
| Pheochromocytoma and Paraganglioma | PCPG | 3 | 179 | 3 | 178 |
| Prostate adenocarcinoma | PRAD | 49 | 340 | 52 | 374 |
| Rectum adenocarcinoma | READ | 7 | 98 | 9 | 91 |
| Sarcoma | SARC | 4 | 242 | 2 | 103 |
| Skin Cutaneous Melanoma | SKCM | 2 | 92 | 1 | 82 |
| Stomach adenocarcinoma | STAD | 2 | 339 | 0 | 0 |
| Thyroid carcinoma | THCA | 56 | 507 | 59 | 498 |
| Uterine Carcinosarcoma | UCS | 0 | 57 | 0 | 57 |
| Uterine Corpus Endometrial Carcinoma | UCEC | 46 | 438 | 24 | 158 |
| Uveal Melanoma | UVM | 0 | 80 | 0 | 0 |
Wanderer offers external links to [Genome Browser](http://genome-euro.ucsc.edu/), [Regulome Explorer](http://explorer.cancerregulome.org/all_pairs/) and [cBioPortal](http://cbioportal.org/). We note that some of the user's queries might involve a gene or a dataset that was not covered in either Regulome Explorer or cBioPortal (i.e. mesothelioma). We note that the TCGA participants used in the analysis might differ between Wanderer and these others Web resources, as they used different data snapshots from TCGA. Moreover, Regulome Explorer datasets might contain tumor-only data (without the normal adjacent).
| Description | Name | Regulome Explorer | cBioPortal |
|----------------------- |:------:|:-----:|:-----:|:-----:|
| Adrenocortical carcinoma | ACC | acc_2015_03_31 | acc_tcga |
| Bladder Urothelial Carcinoma | BLCA | blca_20may13_test | blca_tcga_pub |
| Brain Lower Grade Glioma | LGG | lgg_04oct13_seq | lgg_tcga |
| Breast invasive carcinoma | BRCA | brca_manuscript_rerun_nov12d_pw | brca_tcga_all |
| Cervical squamous cell carcinoma and endocervical adenocarcinoma | CESC | | cesc_tcga |
| Colon adenocarcinoma | COAD| coad_03feb13_seq_tumor_only | coadread_tcga_all |
| Esophageal carcinoma | ESCA | | esca_tcga |
| Glioblastoma multiforme | GBM | gbm_2013_pub_tumor_only | gbm_tcga_all |
| Head and Neck squamous cell carcinoma | HNSC | hnsc_03feb13_seq_tumor_only | hnsc_tcga |
| Kidney Chromophobe | KICH | | kich_tcga_pub |
| Kidney renal clear cell carcinoma | KIRC | kirc_01oct12_A_pw | kirc_tcga_pub |
| Kidney renal papillary cell carcinoma | KIRP | | kirp_tcga |
| Liver hepatocellular carcinoma | LIHC | | lihc_tcga |
| Lung adenocarcinoma | LUAD | luad_03feb13_seq_tumor_only | luad_tcga_all |
| Lung squamous cell carcinoma | LUSC | lusc_03feb13_seq_tumor_only | lusc_tcga |
| Ovarian serous cystadenocarcinoma | OV | ov_03feb13_ary_tumor_only | ov_tcga_pub |
| Pancreatic adenocarcinoma | PAAD | | paad_tcga |
| Pheochromocytoma and Paraganglioma | PCPG | | pcpg_tcga |
| Prostate adenocarcinoma | PRAD | | prad_tcga |
| Rectum adenocarcinoma | READ | | coadread_tcga |
| Sarcoma | SARC | | sarc_tcga |
| Skin Cutaneous Melanoma | SKCM | | skcm_tcga |
| Stomach adenocarcinoma | STAD | stad_23jan14_seq_tumor_only | stad_tcga |
| Thyroid carcinoma | THCA | thca_18oct14_TP | thca_tcga |
| Uterine Carcinosarcoma | UCS | | ucs_tcga |
| Uterine Corpus Endometrial Carcinoma | UCEC | ucec_28jun13b_seq_tumor_only | ucec_tcga |
| Uveal Melanoma | UVM | uvm_20150515_private | | |
### 6. FREQUENTLY ASKED QUESTIONS ###########################################################################
6.1. **Wanderer says there are not enough samples to perform statistical analysis, what does it mean?**
This means there were not enough number of informative samples to compute the Wilcoxon test. We note that, although we calculate the test with small number of samples, this result might be meaningless without a minimum number of cases.
6.2. **The CSV files are a single line file!**
We use Linux/UNIX linefeeds. Although your text viewer (i.e. notepad) might merge all the lines under Windows, your spreadsheet software (i.e. Excel) will properly recognize the line separation.
6.3. **How can I open a CSV in my spreadsheet software?**
In most cases it should work by just using the **open with** (right click with the mouse on the file icon). Alternatively, launch the application and import data from text, comma-separated or character-delimited values (this depends on the software, i.e. LibreOffice).
6.4. **How and when did you download the data from the TCGA?**
Data from TCGA was downloaded using the TCGA-Assembler [(Zhu et al., 2014)][(Zhu et al., 2014)] on July 8th 2014.
6.6. **What is the origin of annotation data?**
450k methylation array probes have been annotated as described in [(Price et al., 2013)][(Price et al., 2013)].
Expression data annotation was performed using gene annotations and exon positions as provided by the TCGA and UCSC Genome Browser. Data was fetched using biomaRt package from R and, when needed, was merged according to feature overlaps [(Gel et al., 2015)].
6.6. **What do RPKM and RSEM stand for?**
Both values are directly obtained from the TCGA pipeline for RNA seq v2 expression analysis. RPKM stands for *reads per kilobase per million mapped reads* [(Guo et al., 2013)] and is used to represent exon expression levels. RSEM stands for *RNA-Seq by Expectation-Maximization* [(Guo et al., 2013)] and is used to represent gene expression levels. We log2-transform these values adding one to get rid of zeroes; that means we plot log_2(x+1) RPKM or RSEM values.
6.7. **Is the Wilcoxon test coupled to the box plots pairwise? For instance, if the breast cancer has n=113 normal and n=1052 tumor how many samples are taken into account for the comparison?**
The test is not a pairwise Wilcoxon one: it does not compare the tumor to the matched normal. It actually tests all the normal data vs all the tumor data, regardless of the pairing. In this example it would compare n=113 against n=1052.
### 7. REFERENCES ############################################################################################
[(Gel et al., 2015)]: http://master.bioconductor.org/packages/release/bioc/html/regioneR.html (regioneR)
[(Gel et al., 2015)] Gel B., Diez-Villanueva A., Serra E., Buschbeck M., Peinado M.A. and Malinverni R. **regioneR: an R/Bioconductor package for the association analysis of genomic regions based on permutation tests.** Under review in Bioinformatics (2015)
[(Guo et al., 2013)]: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0071462 (Large Scale Comparison of Gene Expression Levels by Microarrays and RNAseq Using TCGA Data)
[(Guo et al., 2013)] Guo Y., Sheng Q., Li J., Ye F., Samuels D.C. and Shyr Y. **Large scale comparison of gene expression levels by microarrays and RNAseq using TCGA data.** PloS One (2013), 8:8 doi:10.1371/journal.pone.0071462.
[(Price et al., 2013)]: http://www.epigeneticsandchromatin.com/content/6/1/4 (Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array)
[(Price et al., 2013)] Price M.E., Cotton A.M., Lam L.L., Farré P., Emberly E., Brown C.J., Robinson W.P. and Kobor M.S. **Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array**. Epigenetics & Chromatin (2013), 6:4 doi:10.1186/1756-8935-6-4.
[(Zeeberg et al., 2004)]: http://www.biomedcentral.com/1471-2105/5/80 (Mistaken Identifiers: Gene name errors can be introduced inadvertently when using Excel in bioinformatics)
[(Zeeberg et al., 2004)] Zeeberg .BR., Riss J., Kane D.W., Bussey K.J., Uchio E., Linehan W.M., Barrett J.C. and Weinstein JN. **Mistaken identifiers: gene name errors can be introduced inadvertently when using Excel in bioinformatics.** BMC Bioinformatics (2004) Jun 23;5:80.
[(Zhu et al., 2014)]: http://www.ncbi.nlm.nih.gov/pubmed/24874569 (TCGA-assembler: open-source software for retrieving and processing TCGA data.)
[(Zhu et al., 2014)] Zhu Y., Qiu P., Ji Y. **TCGA-Assembler: Open-Source Software for Retrieving and Processing TCGA Data**. Nature Methods (2014) 11:599-600 doi:10.1038/nmeth.2956.
### 8. VERSIONS AND COPYRIGHT ##############################################################################
This documentation corresponds to **Wanderer v1.1**. The full development log can be found at https://sourceforge.net/projects/tcga-wanderer/. Wanderer is free software and is provided under the GPL v2 terms.
The previous major releases are:
- **Wanderer v1.1** 2017-05-26 (commit 2f04df36aa6413a21781aa694db3f47f87e5975a)
> Wanderer v1.1 fixes unexpected pairings between normal and tumor barcodes when correlating DNA methylation with gene expression
- **Wanderer v1.0** Wed Jun 10 10:56:31 2015 +0200 (commit 9339d9ee59d6fabaf802b1d17caaadbf82786442)
> Fixes minor bugs on profiling, allows data download in a ZIP file, adds new datasets, includes RNA seq by-gene data and correlates methylation and expression.
- **Alpha Release** Mon Dec 29 15:14:48 2014 (commit 21f1c7d9c742011e289842752a1ea871c13dc5fb)
> Profiles DNA methylation and expression for matched normal-tumor datasets.
- **Project start** Mon Oct 6 15:04:45 2014 (commit 3e5440a552d04319331415e74635e144045cc6ae)
> Project prototype.
Some datasets have limitations for usage until a global analysis is published; please contact TCGA before publishing.
Copyright IMPPC, 2014 - Izaskun Mallona, Anna Diez-Villanueva and Miguel A. Peinado. Logo by Julien Douet.
### 9. CONTACT ##############################################################################################
Feel free to contact Izaskun Mallona (izaskun.mallona@gmail.com) or Miguel A. Peinado (mpeinado@igtp.cat).
We provide more tools and data at http://www.maplab.cat.
]]>