In sigven/oncoEnrichR: Cancer-dedicated gene set interpretation

Prognostic associations

Gene expression associations - Human Protein Atlas {.tabset}

Based on data from the Human Protein Atlas - Pathology Atlas, we are here listing significant results from correlation analyses of mRNA expression levels of human genes in tumor tissue and the clinical outcome (survival) for ~8,000 cancer patients (TCGA)
All correlation analyses have been performed in a gene-centric manner, and associations are only shown for genes in the query set. We separate between
Favorable associations : high expression of a given gene is associated with better survival
Unfavorable associations : high expression of a given gene is associated with worse survival
Strength of associations are provided through p-values (only associations with a p-value< = 0.001 are provided), in addition we provide a percentile rank for associations considering a) all significant (p-value <= 0.001) associations across all tumor sites (column percentile_rank_all), and b) only significant (p-value <= 0.001) associations found in the same tumor site (column percentile_rank_site)

show_hpa_prognostic <- list()
prognostic_assoc_data <- list()
active_tab <- list()

for(n in c('favorable','unfavorable')){
    show_hpa_prognostic[[n]] <- TRUE
    active_tab[[n]] <- FALSE
    prognostic_assoc_data[[n]] <- data.frame()

    prognostic_assoc_data[[n]] <- 
      onc_enrich_report[['data']][['cancer_prognosis']][['hpa']][['assocs']] |> 
      dplyr::filter(evidence_direction == n) |>
      dplyr::select(-evidence_direction)

    if(nrow(prognostic_assoc_data[[n]]) == 0){
      show_hpa_prognostic[[n]] <- FALSE
    }
}

if(show_hpa_prognostic[['favorable']] == T){
  active_tab[['favorable']] <- T
}else{
  if(show_hpa_prognostic[['unfavorable']] == T){
    active_tab[['unfavorable']] <- T
  }else{
    active_tab[['favorable']] <- T
  }
}

if(active_tab[['favorable']] == T){
  cat("")
  cat("##### Favorable associations {.active}")
  cat("")
}else{
  cat("")
  cat("##### Favorable associations")
  cat("")
}

prognostic_genes_favorable <- crosstalk::SharedData$new(prognostic_assoc_data[['favorable']])

crosstalk::bscols(
  list(
    crosstalk::filter_select("symbol", "Target gene", prognostic_genes_favorable, ~symbol),
    crosstalk::filter_slider("log10_p_value", "-log10 (p-value)", 
                             prognostic_genes_favorable, ~log10_p_value, step = 1,
                             min = 3, max = 16)
  ),
  list(
    crosstalk::filter_select("primary_site", "Primary tumor site", prognostic_genes_favorable, ~primary_site),
    crosstalk::filter_slider("percentile_rank_all", "Percentile rank (any tumor site)", 
                             prognostic_genes_favorable, ~percentile_rank_all, 
                             step = 10, min = 0, max = 100),
    crosstalk::filter_slider("percentile_rank_site", "Percentile rank (tumor site)", 
                             prognostic_genes_favorable, ~percentile_rank_site, 
                             step = 10, min = 0, max = 100)

  )
)

htmltools::br()
DT::datatable(prognostic_genes_favorable, escape = F, 
              extensions=c("Buttons","Responsive"), width = "100%",
              options=list(buttons = c('csv','excel'), 
                           dom = 'Bfrtip')) |>
  DT::formatStyle("symbol","log10_p_value", 
                  color = "white", 
                  fontWeight = "bold",
                  backgroundColor = 
                    DT::styleInterval(onc_enrich_report[['config']][['prognosis']][['breaks']], onc_enrich_report[['config']][['prognosis']][['colors_favorable']]))

cat('<br><br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;No genes in the query set with favorable gene expression association to prognosis&nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')
cat('<br><br>')

if(active_tab[['unfavorable']] == T){
  cat("")
  cat("##### Unfavorable associations {.active}")
  cat("")
}else{
  cat("")
  cat("##### Unfavorable associations")
  cat("")
}

prognostic_genes_unfavorable <- crosstalk::SharedData$new(prognostic_assoc_data[['unfavorable']])

crosstalk::bscols(
  list(
    crosstalk::filter_select("symbol", "Target gene", prognostic_genes_unfavorable, ~symbol),
    crosstalk::filter_slider("log10_p_value", "-log10 (p-value)", 
                             prognostic_genes_unfavorable, ~log10_p_value, step = 1,
                             min = 3, max = 16)
  ),
  list(
    crosstalk::filter_select("primary_site", "Primary tumor site", prognostic_genes_unfavorable, ~primary_site),
    crosstalk::filter_slider("percentile_rank_all", "Percentile rank (any tumor site)", 
                             prognostic_genes_unfavorable, ~percentile_rank_all, 
                             step = 10, min = 0, max = 100),
    crosstalk::filter_slider("percentile_rank_site", "Percentile rank (tumor site)", 
                             prognostic_genes_unfavorable, ~percentile_rank_site, 
                             step = 10, min = 0, max = 100)

  )
)

htmltools::br()
DT::datatable(prognostic_genes_unfavorable, escape = F, 
              extensions=c("Buttons","Responsive"), width = "100%",
              options=list(buttons = c('csv','excel'), 
                           dom = 'Bfrtip')) |>
  DT::formatStyle("symbol","log10_p_value", 
                  color = "white", 
                  fontWeight = "bold",
                  backgroundColor = 
                    DT::styleInterval(onc_enrich_report[['config']][['prognosis']][['breaks']], onc_enrich_report[['config']][['prognosis']][['colors_unfavorable']]))

cat('<br><br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;No genes in the query set with unfavorable gene expression association to prognosis&nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')
cat('<br><br>')

Genetic determinants of survival in cancer {.tabset}

Based on data recently calculated by Smith et al., Cell Reports, 2022, we here show the relative prognostic implications of genes in the query set, for different genetic features ( here limited to expression, mutation, methylation, and CNA).
The data provided by Smith et al. was harvested through analysis of TCGA datasets, in which Cox proportional hazards models were generated, linking the expression, copy number, methylation, or mutation status of every gene in the genome with patient outcome in different cancer types profiled by the TCGA.
Kaplan-Meier curves were generated by dividing patients into two groups and comparing the survival times between each group.
- For RNA-Seq, the division is made based on the mean expression of the feature.
- For CNAs, the division is made based on the mean copy number of the feature.
- For methylation, the division is made based on the mean methylation level of the feature.
- For mutations, the division is made between patients who have a mutation (protein-coding only) in a gene and patients who lack mutations in a gene.
Each gene has been attributed, for each genetic feature, with a Z-score (Wald statistic), indicating the relative strength of association of a particular feature to either
- Death (positive scores) , or
- Survival (negative scores)
Shown here are four heatmaps (CNA, expression, mutation, methylation) with prognostic Z-scores for members of the query set (limited to the top 100). Tumor type cohorts are designated with TCGA study type abbreviations (e.g. COAD = Colon Adenocarcinoma, BRCA = Breast Invasive Carcinoma etc.).

Mutation

mutation_zscore_matrix <- as.data.frame(
  onc_enrich_report[['data']][['cancer_prognosis']][['km_cshl']][['assocs']][['mut']] |>
  tidyr::pivot_wider(names_from = tcga_cohort,
                     values_from = z_score) |>
    dplyr::arrange(symbol)
)
rownames(mutation_zscore_matrix) <-
  mutation_zscore_matrix$symbol
mutation_zscore_matrix$symbol <- NULL
mutation_zscore_matrix <- as.matrix(mutation_zscore_matrix)

if(NROW(mutation_zscore_matrix) > 100){
  mutation_zscore_matrix <- mutation_zscore_matrix[1:100,]
}

if(NROW(mutation_zscore_matrix) > 1){
  mutation_zscore_matrix <- 
    mutation_zscore_matrix[nrow(mutation_zscore_matrix):1, ]
}

n_zscore_ticks <- 20
if(NROW(mutation_zscore_matrix) < 20){
  n_zscore_ticks <- 5
}

mutation_survival_fig <- plotly::plot_ly(
  colors = "RdBu",
  width = 800,
  height = 400 + (14.67 * NROW(mutation_zscore_matrix))) |>
  plotly::add_heatmap(
    y = rownames(mutation_zscore_matrix),
    x = colnames(mutation_zscore_matrix),
    z = mutation_zscore_matrix,
    hovertext = "Z-score",
    yaxis = "y") |>
  plotly::layout(
    title = 'Target survival associations (mutation)',
    xaxis = list(tickfont = list(size = 13, family = "Helvetica"), 
                 tickangle = -50),
    yaxis = list(tickfont = list(size = 12, family = "Helvetica")),
    margin = list(l = 75, r = 20, b = 150, t = 30, pad = 4)
  ) |>
  plotly::colorbar(
    nticks = n_zscore_ticks, 
    title = list(text = "Z-score",side = "bottom"),
    limits = c(-8.5, 8.5))

mutation_survival_fig

cat('<br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;<b>NO GENES</b> in the query set with survival association scores (mutation). &nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')

Expression

expression_zscore_matrix <- as.data.frame(
  onc_enrich_report[['data']][['cancer_prognosis']][['km_cshl']][['assocs']][['exp']] |>
  tidyr::pivot_wider(names_from = tcga_cohort,
                     values_from = z_score) |>
    dplyr::arrange(symbol)
)
rownames(expression_zscore_matrix) <-
  expression_zscore_matrix$symbol
expression_zscore_matrix$symbol <- NULL
expression_zscore_matrix <- as.matrix(expression_zscore_matrix)

if(NROW(expression_zscore_matrix) > 100){
  expression_zscore_matrix <- expression_zscore_matrix[1:100,]
}

if(NROW(expression_zscore_matrix) > 1){
  expression_zscore_matrix <- 
    expression_zscore_matrix[nrow(expression_zscore_matrix):1, ]
}

n_zscore_ticks <- 20
if(NROW(expression_zscore_matrix) < 20){
  n_zscore_ticks <- 5
}

expression_survival_fig <- plotly::plot_ly(
  colors = "RdBu",
  width = 800,
  height = 400 + (14.67 * NROW(expression_zscore_matrix))) |>
  plotly::add_heatmap(
    y = rownames(expression_zscore_matrix),
    x = colnames(expression_zscore_matrix),
    z = expression_zscore_matrix,
    hovertext = "Z-score",
    yaxis = "y") |>
  plotly::layout(
    title = 'Target survival associations (expression)',
    xaxis = list(tickfont = list(size = 13, family = "Helvetica"), 
                 tickangle = -50),
    yaxis = list(tickfont = list(size = 12, family = "Helvetica")),
    margin = list(l = 75, r = 20, b = 150, t = 30, pad = 4)
  ) |>
  plotly::colorbar(
    nticks = n_zscore_ticks, 
    title = list(text = "Z-score",side = "bottom"),
   limits = c(-11, 11))

expression_survival_fig

cat('<br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;<b>NO GENES</b> in the query set with survival association scores (expression). &nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')

CNA

cna_zscore_matrix <- as.data.frame(
  onc_enrich_report[['data']][['cancer_prognosis']][['km_cshl']][['assocs']][['cna']] |>
  tidyr::pivot_wider(names_from = tcga_cohort,
                     values_from = z_score) |>
    dplyr::arrange(symbol)
)
rownames(cna_zscore_matrix) <-
  cna_zscore_matrix$symbol
cna_zscore_matrix$symbol <- NULL
cna_zscore_matrix <- as.matrix(cna_zscore_matrix)

if(NROW(cna_zscore_matrix) > 100){
  cna_zscore_matrix <- cna_zscore_matrix[1:100,]
}

if(NROW(cna_zscore_matrix) > 1){
  cna_zscore_matrix <- 
    cna_zscore_matrix[nrow(cna_zscore_matrix):1, ]
}

n_zscore_ticks <- 20
if(NROW(cna_zscore_matrix) < 20){
  n_zscore_ticks <- 5
}

cna_survival_fig <- plotly::plot_ly(
  colors = "RdBu",
  width = 800,
  height = 400 + (14.67 * NROW(cna_zscore_matrix))) |>
  plotly::add_heatmap(
    y = rownames(cna_zscore_matrix),
    x = colnames(cna_zscore_matrix),
    z = cna_zscore_matrix,
    hovertext = "Z-score",
    yaxis = "y") |>
  plotly::layout(
    title = 'Target survival associations (CNA)',
    xaxis = list(tickfont = list(size = 13, family = "Helvetica"), 
                 tickangle = -50),
    yaxis = list(tickfont = list(size = 12, family = "Helvetica")),
    margin = list(l = 75, r = 20, b = 150, t = 30, pad = 4)
  ) |>
  plotly::colorbar(
    nticks = n_zscore_ticks, 
    title = list(text = "Z-score",side = "bottom"),
    limits = c(-11, 11))

cna_survival_fig

cat('<br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;<b>NO GENES</b> in the query set with survival association scores (CNA). &nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')

Methylation

meth_zscore_matrix <- as.data.frame(
  onc_enrich_report[['data']][['cancer_prognosis']][['km_cshl']][['assocs']][['meth']] |>
  tidyr::pivot_wider(names_from = tcga_cohort,
                     values_from = z_score) |>
    dplyr::arrange(symbol)
)
rownames(meth_zscore_matrix) <-
  meth_zscore_matrix$symbol
meth_zscore_matrix$symbol <- NULL
meth_zscore_matrix <- as.matrix(meth_zscore_matrix)

if(NROW(meth_zscore_matrix) > 100){
  meth_zscore_matrix <- meth_zscore_matrix[1:100,]
}

if(NROW(meth_zscore_matrix) > 1){
  meth_zscore_matrix <- 
    meth_zscore_matrix[nrow(meth_zscore_matrix):1, ]
}

n_zscore_ticks <- 20
if(NROW(meth_zscore_matrix) < 20){
  n_zscore_ticks <- 5
}

meth_survival_fig <- plotly::plot_ly(
  colors = "RdBu",
  width = 800,
  height = 400 + (14.67 * NROW(meth_zscore_matrix))) |>
  plotly::add_heatmap(
    y = rownames(meth_zscore_matrix),
    x = colnames(meth_zscore_matrix),
    z = meth_zscore_matrix,
    hovertext = "Z-score",
    yaxis = "y") |>
  plotly::layout(
    title = 'Target survival associations (methylation)',
    xaxis = list(tickfont = list(size = 13, family = "Helvetica"), 
                 tickangle = -50),
    yaxis = list(tickfont = list(size = 12, family = "Helvetica")),
    margin = list(l = 75, r = 20, b = 150, t = 30, pad = 4)
  ) |>
  plotly::colorbar(
    nticks = n_zscore_ticks, 
    title = list(text = "Z-score",side = "bottom"),
    limits = c(-11.5, 11.5))

meth_survival_fig

cat('<br>\n <ul><li>  <i> <span style="font-size: 105%; padding: 3px; background-color:#989898; color:white">&nbsp;&nbsp;<b>NO GENES</b> in the query set with survival association scores (methylation). &nbsp;&nbsp; </span></i></li></ul>',sep='\n')
cat('\n')

{.unlisted .unnumbered .toc-ignore}

Citation Note : If you use the output of the Prognostic associations module of oncoEnrichR in your research, please cite the following resources and tools: