--- title: "Cancer Testis explorer" author: - name: Julie Devis, Laurent Gatto, Axelle Loriot package: CTexploreR output: BiocStyle::html_document: toc_float: true df_print : paged vignette: > %\VignetteIndexEntry{Cancer Testis Explorer} %\VignetteEngine{knitr::rmarkdown} %%\VignetteKeywords{Cancer, Testis, Gene expression, methylation, Homo sapiens} %\VignetteEncoding{UTF-8} --- ```{r style, echo = FALSE, results = 'asis'} BiocStyle::markdown() ``` # Introduction Cancer-Testis (CT) genes, also called Cancer-Germline (CG), are a group of genes whose expression is normally restricted to the germline but that are found aberrantly activated in many types of cancers. These genes produce cancer specific antigens represent ideal targets for anti-cancer vaccines. Besides their interest in immunotherapy, they can also be used as cancer biomarkers and as target of anti-tumor therapies with limited side effect. Many CT genes use DNA methylation as a primary mechanism of transcriptionnal regulation. This is another interesting point about CT genes as they represent suitable models to study DNA demethylation in cancer. Currently the reference database gathering CT genes is the [CTdatabase](http://www.cta.lncc.br/) that was published in 2009, based on a literature screening(Almeida et al., 2009). This database is however not up to date. Recently identified CT genes are not referenced (in particular CT genes identified by omics methods that didn’t exist at the time) while some genes referred as CT genes appeared to be in reality expressed in many somatic tissues. Furthermore, the database is not in an easily importable format, some genes are not encoded properly (by synonyms names rather than by their official HGNC symbol names, or by a concatenation of both) resulting in poor interoperability for downstream analyses. More recent studies proposed other lists of CT genes like Wang’s CTatlas (Wang et al., 2016, Jamin et al., 2021, Carter et al., 2023). These lists were established using different criteria to define CT genes and hence differ substantially from each other. Moreover, these lists are usually provided as supplemental data files and are not strictly speaking databases. Finally, none of these studies describe the involvement of DNA methylation in the regulation of individual CT genes. We therefore created `CTexploreR`, a Bioconductor R package, aimed to redefine rigorously the list of CT genes based on publicly available RNAseq databases and to summarize their main characteristics. We included methylation analyses to classify these genes according to whether they are regulated or not by DNA methylation. The package also offers tools to visualize CT genes expression and promoter DNA methylation in normal and tumoral tissues. CTexploreR hence represents an up-to-date reference database for CT genes and can be used as a starting point for further investigations related to these genes. # Installation To install the package: ```{r install1, eval = FALSE} if (!require("BiocManager")) { install.packages("BiocManager") } BiocManager::install("CTexploreR") ``` To install the package from GitHub: ```{r install2, eval = FALSE} if (!require("BiocManager")) { install.packages("BiocManager") } BiocManager::install("UCLouvain-CBIO/CTexploreR") ``` # CT genes The central element of `CTexploreR` is the list of `r nrow(CTexploreR::CT_genes)` CT and CTP genes (see table below) selected based on their expression in normal and tumoral tissues (selection details in the next section). The table also summarises their main characteristics. ```{r CT_genes} library(CTexploreR) head(CT_genes, 10) ``` `CTdata` is the companion Package for `CTexploreR` and provides the omics data that was necessary to select and characterize cancer testis genes as well as exploring them. The data are served through the `ExperimentHub` infrastructure. Currently available data are summarised in the table below and details can be found in `CTdata` vignette or manuals. ```{r data} CTdata() ``` # CT gene selection In order to generate the list of CT genes, we followed a specific selection procedure (see figure below). ```{r ctselection, results='markup', echo=FALSE, fig.align='center', out.width = '100%'} knitr::include_graphics("./figs/Figure_CT.png") ``` ## Testis-specific expression Testis-specific genes (expressed exclusively in testis) and testis-preferential genes (expressed in a few somatic tissues at a level lower than 10x testis expression) were first selected using the GTEx database (Aguet et al., 2020). Note that some genes were undetectable in the GTEx database due to multimapping issues (these were flagged as "lowly expressed" in `GTEX_category` column). A careful inspection of these genes showed that many of them are well-known Cancer-Testis genes, belonging to gene families (MAGEA, SSX, CT45A, GAGE, …) from which members have identical or nearly identical sequences. This is likely the reason why these genes are not detected in the GTEx database, as GTEx processing pipeline specifies that overlapping intervals between genes are excluded from all genes for counting.For these genes, as testis-specificity could not be assessed using GTEx database, RNAseq data from a set of normal tissues were reprocessed in order to allow multimapping. Expression of several genes became detectable in some tissues. Genes showing a testis-specific expression (expression at least 10x higher in testis than in any somatic tissues when multimapping was allowed) were selected, and flagged as testis-specific in `multimapping_analysis` column. Additionally, as our selection procedure is based on bulk RNAseq data, we wanted to ensure that the selected genes are not expressed in rare populations of somatic cells. We used the Single Cell Type Atlas classification from the Human Protein Atlas (Uhlén et al., 2015) to exclude the ones that were flagged as specific of any somatic cell type. ## Activation in cancer cell lines and TCGA tumors To assess activation in cancers, RNAseq data from cancer cell lines from CCLE (Barretina et al., 2012) and from TCGA cancer samples (Cancer Genome Atlas Research Network et al., 2013) were used. This allowed to select among testis-specific and testis-preferential genes those that are activated in cancers. In the `CCLE_category` and `TCGA_category` columns, genes are tagged as "activated" when they are highly expressed in at one percent of cancer cell line/sample (TPM >= 1). However genes that were found to be expressed in all -or almost all-cancer cell lines/samples were removed, as this probably reflects a constitutive expression rather than a true activation. We filtered out genes that were not completely repressed in at least 20 % of cancer cell lines/samples (TPM <= 0.5). We also made use of the normal peritumoral samples available in TCGA data to remove from our selection genes that were already detected in a significant fraction of these cells. ## IGV visualisation All selected CT genes were visualised on IGV (Thorvaldsdóttir et al., 2013) using a RNA-seq alignment from testis, to ensure that expression in testis really corresponded to the canonical transcript. The aim was initially to identify precisely the transcription start site of each gene, but unexpectedly we observed that for some genes, the reads were not properly aligned on exons, but were instead spread across a wide genomic region spanning the genes. These genes, flagged as "unclear" in `IGV_backbone` column, were removed from the CT_gene category, as their expression values in GTEX, TCGA and CCLE might reflect a poorly defined transcription in these regions and are hence likely unreliable. ## Regulation by methylation Genes flagged as `TRUE` in `regulated_by_methylation` column correspond to * Genes that are significantly induced by a demethylating agent (RNAseq analysis of cell lines reated with DAC (5-Aza-2′-Deoxycytidine)). * Genes that have a highly methylated promoter in normal somatic tissues (WGBS analysis of a set of normal tissues). For some genes showing a strong activation in cells treated with 5-Aza-2′-Deoxycytidine, methylation analysis was not possible due to multimapping issues. In this case, genes were still considered as regulated by methylation unless their promoter appeared unmethylated in somatic tissues or methylated in germ cells. # Available functions For details about functions, see their respective manual pages. For all functions, an option `values_only` can be set to `TRUE` in order to get the values instead of the visualisation. All visualisation functions can be used on all GTEx genes, not only on Cancer-Testis genes, as the data they refer to contains all genes. By default, if no genes are specified in a function, only strict CT genes will be used. If one wants all 280 CT genes to be used, an option `include_CTP` can be set to `TRUE` to also use CT preferential genes. ## Expression in normal healthy adult tissues ### `GTEX_expression()` Allows to visualise gene expression in GTEx tissues. We can for example see the difference of expression between testis-specific and testis-preferential genes. Testis-specific genes have been determined with a stricter specificity to the testis : they are lowly expressed in all somatic tissues and at least 10 times more in the testis. Whereas testis-preferential accepts a little expression outside the testis : they are lowly expressed in at least 75% of somatic tissues, but still 10 times more in the testis. * Applied to testis-specific genes : we can clearly see the expression strictly limited to the testis. We can also see genes that are lowly expressed in GTEx, and have thus been characterized using multimapping (see below). ```{r} testis_specific <- dplyr::filter( CT_genes, CT_gene_type == "CT_gene") GTEX_expression(testis_specific$external_gene_name, units = "log_TPM") ``` * Applied to testis-preferential genes : we can see that the expression is less stringent to testis, as expected, with low expression in some other tissues, always with a strong testis signal. ```{r} testis_preferential <- dplyr::filter( CT_genes, CT_gene_type == "CTP_gene") GTEX_expression(testis_preferential$external_gene_name, units = "log_TPM") ``` ### `normal_tissue_expression_multimapping()` Allows to visualise the expression values obtained by counting or not multi-mapped reads in normal tissues. We can apply that to genes that are lowly expressed in GTEx (as seen above) in order to visualise their expression. First heatmap shows expression without multimapping while multimapping was allowed in the second one, where expression can be observed. This allowed their determination as testis-specific genes. ```{r} testis_specific_in_multimapping_analysis <- dplyr::filter(CT_genes, lowly_expressed_in_GTEX) normal_tissue_expression_multimapping( testis_specific_in_multimapping_analysis$external_gene_name, multimapping = FALSE, units = "log_TPM") normal_tissue_expression_multimapping( testis_specific_in_multimapping_analysis$external_gene_name, multimapping = TRUE, units = "log_TPM") ``` ### `testis_expression()` Allows to visualise gene expression in all different testis cell type, somatic or germ, using data from the adult human testis transcriptional cell atlas. Using CT genes localisation on the X chromosome, we can see that they tend to be expressed in th early stages of spermatogenesis when located on the X. We also visualise clearly that genes mainly expressed in an early stage of spermatogenesis aren't expressed later and vice-versa. ```{r} X_CT <- dplyr::filter(CT_genes, X_linked) testis_expression(X_CT$external_gene_name, cells = "germ_cells") notX_CT <- dplyr::filter(CT_genes, !X_linked) testis_expression(notX_CT$external_gene_name, cells = "germ_cells") ``` ### `oocytes_expression()` Allows to visualise gene expression oocytes at different stage, using scRNA-seq data form "Decoding dynamic epigenetic landscapes in human oocytes using single-cell multi-omics sequencing" (Yan et al. Cell Stem Cell 2021) We can here again compare the expression of CT genes that are located or not on the X chromosome. Showing a lesser expression of X-linked genes in oocytes. ```{r} oocytes_expression(X_CT$external_gene_name) oocytes_expression(notX_CT$external_gene_name) ``` ### `HPA_cell_type_expression()` Allows to visualise gene expression in all different healthy cell type, comparing different somatic cells to germ cells, using data from the human proteome atlas. Visualising all CT_genes, the specificity to germ cells only is quite clear. ```{r} HPA_cell_type_expression(units = "scaled") ``` ## Expression in fetal cells No analysis was added to the `all_genes` and `CT_genes` tables using these datasets. ### `embryo_expression()` Allows to visualise gene expression in human early embryos using two different scRNA-seq datasets. * Single-Cell RNA-Seq Reveals Lineage and X Chromosome Dynamics in Human Preimplantation Embryos, Petropoulos et al., Cell 2016 This dataset contains different stages of blastocysts and morula. ```{r} embryo_expression(dataset = "Petropoulos", include_CTP = FALSE) ``` * Single-cell DNA methylome sequencing of human preimplantation embryos, Zhu et al. Nat genetics 2018 This dataset only contains blastocyst stage. ```{r} embryo_expression(dataset = "Zhu", include_CTP = FALSE) ``` ### `fetal_germcells_expression()` Allows to visualise gene expression in fetal germ cells the scRNA-seq dataset from "Single-cell roadmap of human gonadal development" (Garcia-Alonso, Nature 2022). ```{r} fetal_germcells_expression(include_CTP = FALSE, ncells_max = 100) ``` ### `hESC_expression()` Allows to visualise gene expression in human embryonic stem cells, using scRNAseq data downloaded from Encode database. ```{r} hESC_expression(include_CTP = FALSE, units = "log_TPM", values_only = FALSE) ``` ## Expression in cancer cells and samples ### `CCLE_expression()` Allows to visualise gene expression in different histological types of CCLE cancer cell lines. We can thus compare genes that are or not activated in tumoral cell lines. * Applied to CT genes frequently activated in CCLE cell lines, more than 5% of CCLE cell_lines are expressing each gene. A string signal is visible. ```{r} frequently_activated <- dplyr::filter( CT_genes, percent_of_positive_CCLE_cell_lines >= 5) CCLE_expression( genes = frequently_activated$external_gene_name, type = c( "lung", "skin", "bile_duct", "bladder", "colorectal", "lymphoma", "uterine", "myeloma", "kidney", "pancreatic", "brain", "gastric", "breast", "bone", "head_and_neck", "ovarian", "sarcoma", "leukemia", "esophageal", "neuroblastoma"), units = "log_TPM") ``` * Applied to CT genes not frequently activated in CCLE cell lines, more than 95% of cell lines are not expressing each genes. The lack of expression is very clear. ```{r} not_frequently_activated <- dplyr::filter( CT_genes, percent_of_negative_CCLE_cell_lines >= 95) CCLE_expression( genes = not_frequently_activated$external_gene_name, type = c( "lung", "skin", "bile_duct", "bladder", "colorectal", "lymphoma", "uterine", "myeloma", "kidney", "pancreatic", "brain", "gastric", "breast", "bone", "head_and_neck", "ovarian", "sarcoma", "leukemia", "esophageal", "neuroblastoma"), units = "log_TPM") ``` ### `CT_correlated_genes()` A function that use expression data from all CCLE cell lines and returns genes correlated (or anti-correlated) with specified CT gene. Here with MAGEA3, we can see all genes (CT or not) whose expression is correlated with it, like MAGEA6. ```{r} CT_correlated_genes("MAGEA3", 0.3) ``` ### `TCGA_expression()` Allows to visualise gene expression in cancer samples from TCGA. We can thus for example compare with activation in CCLE cell lines. * Applied to CT genes frequently activated in CCLE cell lines ```{r} TCGA_expression( genes = frequently_activated$external_gene_name, tumor = "all", units = "log_TPM") ``` * Applied to CT genes not frequently activated in CCLE cell lines ```{r} TCGA_expression( genes = not_frequently_activated$external_gene_name, tumor = "all", units = "log_TPM") ``` When visualising only one tumor, there is a separation between tumoral and peritumoral tissue. ```{r} TCGA_expression( genes = frequently_activated$external_gene_name, tumor = "LUAD", units = "log_TPM") ``` ## Methylation analysis ### `DAC_induction()` Allows to visualise gene induction upon DAC treatment in a series of cell lines. We can see the difference between CT genes regulated by methylation or not * Applied to genes controlled by methylation ```{r} controlled_by_methylation <- dplyr::filter(CT_genes, regulated_by_methylation) DAC_induction(genes = controlled_by_methylation$external_gene_name) ``` * Applied to genes not controlled by methylation ```{r} not_controlled_by_methylation <- dplyr::filter( CT_genes, !regulated_by_methylation) DAC_induction(genes = not_controlled_by_methylation$external_gene_name) ``` ### `normal_tissues_methylation()` Gives the methylation level of all CpGs located in each promoter region (by default defined as 1000 nt upstream TSS and 200 nt downstream TSS) in a set of normal tissues. Can be used to analyse more precisely the methylation of a particular promoter in normal tissues as methylation values are given CpG by CpG. * Applied to a gene controlled by methylation ```{r} normal_tissues_methylation("MAGEB2") ``` * Applied to a gene not controlled by methylation ```{r} normal_tissues_methylation("LIN28A") ``` ### `normal_tissues_mean_methylation()` Gives the mean methylation level of CpGs located in each promoter region (by default defined as 1000 nt upstream TSS and 200 nt downstream TSS) in a set of normal tissues. When comparing genes controlled or not by methylation, we can see a specific methylation pattern in the first group. * Applied to genes controlled by methylation ```{r} normal_tissues_mean_methylation( genes = controlled_by_methylation$external_gene_name) ``` * Applied to genes not controlled by methylation ```{r} normal_tissues_mean_methylation( genes = not_controlled_by_methylation$external_gene_name) ``` ### `embryos_mean_methylation()` Gives the mean methylation level of CpGs located in each promoter region of any genes in early embryos, using WGSB data from ("Single-cell DNA methylome sequencing of human preimplantation embryos". Zhu et al. Nat genetics 2018). Methylation levels in tissues correspond to the mean methylation of CpGs located in range of 1000 pb upstream and 500 pb downstream from gene TSS. Can be used to explore well known CT genes methylation in embryos as below. ```{r} embryos_mean_methylation(c("MAGEA1", "MAGEA3", "MAGEA4", "MAGEC2", "MAGEB16"), stage = c( "MII Oocyte", "Sperm", "Zygote", "2-cell", "4-cell", "8-cell", "Morula")) ``` ### `fetal_germcells_mean_methylation()` Allows to visualise mean promoter methylation levels ofany genes in fetal germ cells, using WGSB data from "Dissecting the epigenomic dynamics of human fetal germ cell development at single-cell resolution" (Li et al. 2021). Methylation levels in tissues correspond to the mean methylation of CpGs located in range of 1000 pb upstream and 500 pb downstream from gene TSS. Can be used to explore well known CT genes methylation in fetal germ cells, as below. ```{r} fetal_germcells_mean_methylation(c("MAGEA1", "MAGEA3", "MAGEA4", "MAGEC2")) ``` ### `hESC_mean_methylation()` Allows to visualise mean promoter methylation levels of any genes in human embryonic cell lines. WGBS methylation data was downloaded from Encode. Methylation levels in tissues correspond to the mean methylation of CpGs located in range of 1000 pb upstream and 200 pb downstream from gene TSS. Here used on strict CT genes. ```{r} hESC_mean_methylation() ``` ### `TCGA_methylation_expression_correlation()` Shows the correlation between gene expression and promoter methylation in TCGA samples. * Applied to a gene controlled by methylation ```{r} TCGA_methylation_expression_correlation( tumor = "all", gene = "TDRD1") ``` * Applied to a gene not controlled by methylation ```{r} TCGA_methylation_expression_correlation( tumor = "all", gene = "LIN28A") ``` # Interactive heatmaps CTexploreR's heatmaps are compatible with the `InteractiveComplexHeatmap` package. This allows the navigate in the last heatmap generated, search a specific gene as well as creating subsets to facilitate data exploration. More info about `InteractiveComplexHeatmap` available [here](https://jokergoo.github.io/InteractiveComplexHeatmap/). ```{r, eval = FALSE} BiocManager::install("InteractiveComplexHeatmap") library(InteractiveComplexHeatmap) GTEX_expression(testis_specific$external_gene_name, units = "log_TPM") htShiny() ``` # Bibliography Aguet, F., Anand, S., Ardlie, K. G., Gabriel, S., Getz, G. A., Graubert, A., Hadley, K., Handsaker, R. E., Huang, K. H., Kashin, S., Li, X., MacArthur, D. G., Meier, S. R., Nedzel, J. L., Nguyen, D. T., Segrè, A. V., Todres, E., Balliu, B., Barbeira, A. N., … Volpi, S. (2020). The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. https://science.sciencemag.org/content/369/6509/1318.abstract Almeida, L. G., Sakabe, N. J., deOliveira, A. R., Silva, M. C. C., Mundstein, A. S., Cohen, T., Chen, Y.-T., Chua, R., Gurung, S., Gnjatic, S., Jungbluth, A. A., Caballero, O. L., Bairoch, A., Kiesler, E., White, S. L., Simpson, A. J. G., Old, L. J., Camargo, A. A., & Vasconcelos, A. T. R. (2009). CTdatabase: a knowledge-base of high-throughput and curated data on cancer-testis antigens. Nucleic Acids Research, 37(Database issue), D816–D819. Barretina, J., Caponigro, G., Stransky, N., Venkatesan, K., Margolin, A. A., Kim, S., Wilson, C. J., Lehár, J., Kryukov, G. V., Sonkin, D., Reddy, A., Liu, M., Murray, L., Berger, M. F., Monahan, J. E., Morais, P., Meltzer, J., Korejwa, A., Jané-Valbuena, J., … Garraway, L. A. (2012). The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature, 483(7391), 603–607. Cancer Genome Atlas Research Network, Weinstein, J. N., Collisson, E. A., Mills, G. B., Shaw, K. R. M., Ozenberger, B. A., Ellrott, K., Shmulevich, I., Sander, C., & Stuart, J. M. (2013). The Cancer Genome Atlas Pan-Cancer analysis project. Nature Genetics, 45(10), 1113–1120. Carter, J. A., Matta, B., Battaglia, J., Somerville, C., Harris, B. D., LaPan, M., Atwal, G. S., & Barnes, B. J. (2023). Identification of pan-cancer/testis genes and validation of therapeutic targeting in triple-negative breast cancer: Lin28a- and Siglece-based vaccination induces anti-tumor immunity and inhibits metastasis. In bioRxiv (p. 2023.05.09.539617). Guo, J. et al. The adult human testis transcriptional cell atlas. Cell Res. 28, 1141–1157 (2018). Uhlén, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 1260419 (2015). Thorvaldsdóttir, H., Robinson, J. T., & Mesirov, J. P. (2013). Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Briefings in Bioinformatics, 14(2), 178–192. Wang, C., Gu, Y., Zhang, K., Xie, K., Zhu, M., Dai, N., Jiang, Y., Guo, X., Liu, M., Dai, J., Wu, L., Jin, G., Ma, H., Jiang, T., Yin, R., Xia, Y., Liu, L., Wang, S., Shen, B., … Hu, Z. (2016). Systematic identification of genes with a cancer-testis expression pattern in 19 cancer types. Nature Communications, 7, 10499. # Session information {-} ```{r sessioninfo, echo=FALSE} sessionInfo() ```