RGMQL Example R Notebook: Use case 1
Silvia Cascianelli
2021-07-22
In this example we investigate the TCGA mutational data, extracted from the GMQL remote repository, to evaluate the most mutated gene regions in patients affected by Kidney Renal Clear Cell Carcinoma and younger than 65 years.
Load the RGMQL package and initialize the remote GMQL context of scalable data management engine, specifying remote_processing = TRUE, and, possibly, an authenticated login:
library(RGMQL)## Caricamento del pacchetto richiesto: RGMQLlib## GMQL successfully loadedremote_url <- "http://www.gmql.eu/gmql-rest"
init_gmql( url = remote_url, remote_processing = TRUE) # , username = 'XXXX', password = 'XXXX')## [1] "your Token is 4a7e8059-9cae-416f-8aa2-6594a6749c03"Download and extract the list of datasets in the curated remote repository and focus on those concerning mutation events:
dataset_list <- show_datasets_list(remote_url)
list <- unlist(lapply(dataset_list[["datasets"]], function(x) x$name))
grep(pattern = 'mutation', x = list, value = TRUE)## [1] "GRCh38_TCGA_somatic_mutation_masked"
## [2] "GRCh38_TCGA_somatic_mutation_masked_2018_12"
## [3] "GRCh38_TCGA_somatic_mutation_masked_2019_10"Choose one dataset of interest and explore its schema and metadata:
schema<-show_schema(remote_url,datasetName = "public.GRCh38_TCGA_somatic_mutation_masked_2019_10")
all_metadata <- show_all_metadata(dataset = "public.GRCh38_TCGA_somatic_mutation_masked_2019_10")Once the dataset is chosen, read it together with the dataset containing the RefSeq gene annotations of the GRCh38 reference genome:
GRCh38_TCGA_mut <- read_gmql(dataset = "public.GRCh38_TCGA_somatic_mutation_masked_2019_10", is_local = FALSE)
RefSeq_GRCh38 <- read_gmql(dataset = "public.GRCh38_ANNOTATION_REFSEQ", is_local = FALSE)Filter GRCh38_TCGA_mut based on a metadata predicate to keep Kidney Renal Clear Cell Carcinoma mutations only, and based on patient age:
mut_under65 <- filter(GRCh38_TCGA_mut, clinical__clin_shared__age_at_initial_pathologic_diagnosis < 65 &
biospecimen__admin__disease_code == "KIRC")Filter out RefSeq_GRCh38 based on a metadata predicate to keep RefSeq gene regions only:
genes = filter(RefSeq_GRCh38, annotation_type == 'gene' &
provider == 'RefSeq')For each mutation sample, map the mutations to the involved genes using the map() function and count mutations within each gene region storing automatically the value in the default region attribute renamed as ‘mut_count’:
geneMut_data1_under65 <- map(genes, mut_under65, count_name = "mut_count")In each sample, remove the genes that do not contain mutations:
geneMut_data2_under65 <- filter(geneMut_data1_under65, r_predicate = mut_count >= 1)Using the extend() function, count how many genes remain in each sample and store the result as a new attribute named ’geneMut_count’:
geneMut_data3_under65 <- extend(geneMut_data2_under65, geneMut_count = COUNT())Order samples in descending order of geneMut_count:
geneMut_data_res_under65 = arrange(geneMut_data3_under65, list(DESC("geneMut_count")))Launch the remote processing execution to materialize resulting datasets:
collect(geneMut_data_res_under65, name = "geneMut_data_res_under65")
JOB<-execute()Monitor the job status:
trace_job(remote_url, JOB$id)Once the job status is ‘SUCCESS’ download the resulting datasets obtained remotely in the working directory of the local File System:
name_dataset_under <- JOB$datasets[[1]]$name
download_dataset(remote_url, name_dataset_under, path='./Results_use_case_1')Download the resulting datasets as GRangesList objects also in the current R environment:
grl_mut_under_D <- download_as_GRangesList(remote_url, name_dataset_under)Log out from remote engine:
logout_gmql(remote_url)## [1] "Logout"Compute the main statistics related to the geneMut_count values distribution:
summary(as.numeric(grl_mut_under_D@unlistData@elementMetadata@listData[["mut_count"]]))## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.000 1.000 1.000 1.023 1.000 3.000quantile(as.numeric(grl_mut_under_D@unlistData@elementMetadata@listData[["mut_count"]]), c(.90, .95, .99))## 90% 95% 99%
## 1 1 2Plot distributions of geneMut_count, i.e. number of mutated genes per patient, and compute main statistics:
meta_lst <- grl_mut_under_D@metadata
mut_lst <- lapply(meta_lst, function(x) as.numeric(x[["geneMut_count"]]))
mut_ord <- mut_lst[order(unlist(mut_lst), decreasing = FALSE)]
library(ggplot2)
p1 <- plot(x = seq(1,5*length(mut_ord), 5), mut_ord, xlab = '', ylab= 'Number of mutated genes', xaxt = "n")
title(xlab="Patient samples", line=-1, cex.lab=1)
xtick <- seq(1, 5*length(mut_ord), 5)
xlabels <- names(mut_ord)
axis(side = 1, at = xtick, labels = xlabels, las = 2, cex.axis = 0.5)
summary(unlist(mut_lst))## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 6 26 37 39 45 466library(GenomicRanges)## Caricamento del pacchetto richiesto: stats4## Caricamento del pacchetto richiesto: BiocGenerics## Caricamento del pacchetto richiesto: parallel##
## Caricamento pacchetto: 'BiocGenerics'## I seguenti oggetti sono mascherati da 'package:parallel':
##
## clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
## clusterExport, clusterMap, parApply, parCapply, parLapply,
## parLapplyLB, parRapply, parSapply, parSapplyLB## I seguenti oggetti sono mascherati da 'package:stats':
##
## IQR, mad, sd, var, xtabs## I seguenti oggetti sono mascherati da 'package:base':
##
## anyDuplicated, append, as.data.frame, basename, cbind, colnames,
## dirname, do.call, duplicated, eval, evalq, Filter, Find, get, grep,
## grepl, intersect, is.unsorted, lapply, Map, mapply, match, mget,
## order, paste, pmax, pmax.int, pmin, pmin.int, Position, rank,
## rbind, Reduce, rownames, sapply, setdiff, sort, table, tapply,
## union, unique, unsplit, which.max, which.min## Caricamento del pacchetto richiesto: S4Vectors##
## Caricamento pacchetto: 'S4Vectors'## I seguenti oggetti sono mascherati da 'package:base':
##
## expand.grid, I, unname## Caricamento del pacchetto richiesto: IRanges##
## Caricamento pacchetto: 'IRanges'## Il seguente oggetto è mascherato da 'package:grDevices':
##
## windows## Caricamento del pacchetto richiesto: GenomeInfoDbfinal <- list()
for (n in names(grl_mut_under_D@metadata)){ #for each of the 217 Granges
gr <- grl_mut_under_D[[n]]
final[[n]] <- data.frame('gene' = gr@elementMetadata@listData[["gene_symbol"]], 'mut_count' = as.numeric(gr@elementMetadata@listData[["mut_count"]]), 'width' = width(ranges(grl_mut_under_D[[n]])))
}Compute and plot the distributions of all the mutational events occurring on each patient:
tot_mut_pat <- list()
max_mut_pat <- list()
mut_pat <- sapply(final, '[[', 'mut_count')
for (i in 1:length(mut_pat)){
tot_mut_pat[[i]] <- sum(as.numeric(mut_pat[[i]]))
max_mut_pat[[i]] <-max(as.numeric(mut_pat[[i]]))
}
names(tot_mut_pat) <- names(mut_pat)
tot_mut_pat_ord <- tot_mut_pat[order(unlist(tot_mut_pat), decreasing = FALSE)]
library(ggplot2)
p2 <- plot(x = seq(1,5*length(tot_mut_pat_ord), 5), tot_mut_pat_ord, xlab = '', ylab= 'Number of mutational events', xaxt = "n")
title(xlab="Patient samples", line = -1, cex.lab = 1)
xtick <- seq(1, 5*length(mut_ord), 5)
xlabels <- names(tot_mut_pat_ord)
axis(side = 1, at = xtick, labels = xlabels, las = 2, cex.axis = 0.5)
genes <- unlist(sapply(final, '[[', "gene"))
lengths <- unlist(sapply(final, '[[', "width"))
mapping <- data.frame('gene' = genes, 'gene_length' = lengths, row.names = NULL)
all_genes <- unique(genes)
all_genes_len <- mapping[which(!duplicated(mapping[,1])),]
gene_m <- matrix(nrow = length(all_genes), ncol = 6)
gene_df <- data.frame(gene_m)
colnames(gene_df) <- c('Gene', 'Length', 'Mutated_Patients', 'Mutation_counts', 'Mutation_counts_norm', 'Mutation_counts_div_len')
gene_df[,1] <- all_genes_len[,1]
gene_df[,2] <- all_genes_len[,2]
for (i in 1:length(all_genes)){ #for each gene
counter_pt <- 0
counter_mut <- 0
counter_mut_norm_pat <- 0
counter_mut_norm_len <- 0
for(j in 1:length(final)){ #for each patient
f <- final[[j]]
if (gene_df[i,1] %in% f[["gene"]]){
counter_pt <- counter_pt+1
counter_mut <- counter_mut + as.numeric(f[["mut_count"]][which(f[["gene"]] == gene_df[i,1])[1]]) #[1] to avoid issues when there are multiple regions of the same gene and duplicated symbols
counter_mut_norm_len <- counter_mut / gene_df[i,2]
counter_mut_norm_pat <- counter_mut_norm_pat + as.numeric(f[["mut_count"]][which(f[["gene"]] == gene_df[i,1])[1]]) / tot_mut_pat[[j]]
}
gene_df[i,3] <- counter_pt
gene_df[i,4] <- counter_mut
gene_df[i,5] <- counter_mut_norm_pat
gene_df[i,6] <- counter_mut_norm_len
}
}Compute and plot the distributions of all the mutational events across patients occurring for top mutated genes:
gene_df <- gene_df[order(gene_df$Mutation_counts, decreasing=TRUE),]
summary(gene_df)## Gene Length Mutated_Patients Mutation_counts
## Length:4775 Min. : 65 Min. : 1.000 Min. : 1.000
## Class :character 1st Qu.: 14389 1st Qu.: 1.000 1st Qu.: 1.000
## Mode :character Median : 38173 Median : 1.000 Median : 1.000
## Mean : 89143 Mean : 1.766 Mean : 1.806
## 3rd Qu.: 95003 3rd Qu.: 2.000 3rd Qu.: 2.000
## Max. :2473595 Max. :103.000 Max. :107.000
## Mutation_counts_norm Mutation_counts_div_len
## Min. :0.002037 Min. :4.040e-07
## 1st Qu.:0.021277 1st Qu.:1.611e-05
## Median :0.030714 Median :3.879e-05
## Mean :0.045331 Mean :1.700e-04
## 3rd Qu.:0.052632 3rd Qu.:9.927e-05
## Max. :3.084954 Max. :2.247e-02library(ggplot2)
p<-plot(x=seq(1:20), y = gene_df[1:20,4], main='Mutation counts in top mutated genes', ylab = 'Number of mutations across patient samples', xlab = NA, xaxt="n")
xtick <- seq(1, 20)
axis(side = 1, at = xtick, labels = gene_df[1:20,1], las = 2, cex.axis = 0.75)
Compute and plot the number of mutational events across patients occurring for top mutated genes, ordering genes by gene length:
library(ggplot2)
p <- plot(x = gene_df[1:20,2], y = gene_df[1:20,4], ylab = 'Number of mutations across samples', xlab = '', main='Mutation counts compared to gene lengths in top mutated genes', xaxt="n") #xlab = 'Genes'
title(xlab="Gene length", line = -1, cex.lab = 1)
xtick <- gene_df[1:20,2]
axis(side = 1, at = xtick, labels = gene_df[1:20,1], las = 2, cex.axis = 0.75)
Compute and plot the top genes based on mutation count per gene length:
gene_df <- gene_df[order(gene_df$Mutation_counts_div_len, decreasing=TRUE),] #descending geneMut_count_order
summary(gene_df)## Gene Length Mutated_Patients Mutation_counts
## Length:4775 Min. : 65 Min. : 1.000 Min. : 1.000
## Class :character 1st Qu.: 14389 1st Qu.: 1.000 1st Qu.: 1.000
## Mode :character Median : 38173 Median : 1.000 Median : 1.000
## Mean : 89143 Mean : 1.766 Mean : 1.806
## 3rd Qu.: 95003 3rd Qu.: 2.000 3rd Qu.: 2.000
## Max. :2473595 Max. :103.000 Max. :107.000
## Mutation_counts_norm Mutation_counts_div_len
## Min. :0.002037 Min. :4.040e-07
## 1st Qu.:0.021277 1st Qu.:1.611e-05
## Median :0.030714 Median :3.879e-05
## Mean :0.045331 Mean :1.700e-04
## 3rd Qu.:0.052632 3rd Qu.:9.927e-05
## Max. :3.084954 Max. :2.247e-02library(ggplot2)
p <- plot(x = seq(1:20), y = gene_df[1:20,6], main = 'Top mutated genes based on mutation count divided by gene length', ylab = 'Mutation counts per gene length', xlab = NA, xaxt = "n") #xlab = 'Genes'
xtick <- seq(1:20)
axis(side = 1, at = xtick, labels = gene_df[1:20,1], las = 2, cex.axis = 0.7)
Compute and plot the percentages of patients having mutations on top mutated genes:
gene_df <- gene_df[order(gene_df$Mutated_Patients, decreasing=TRUE),] #descending gene_mut_count_order
summary(gene_df)## Gene Length Mutated_Patients Mutation_counts
## Length:4775 Min. : 65 Min. : 1.000 Min. : 1.000
## Class :character 1st Qu.: 14389 1st Qu.: 1.000 1st Qu.: 1.000
## Mode :character Median : 38173 Median : 1.000 Median : 1.000
## Mean : 89143 Mean : 1.766 Mean : 1.806
## 3rd Qu.: 95003 3rd Qu.: 2.000 3rd Qu.: 2.000
## Max. :2473595 Max. :103.000 Max. :107.000
## Mutation_counts_norm Mutation_counts_div_len
## Min. :0.002037 Min. :4.040e-07
## 1st Qu.:0.021277 1st Qu.:1.611e-05
## Median :0.030714 Median :3.879e-05
## Mean :0.045331 Mean :1.700e-04
## 3rd Qu.:0.052632 3rd Qu.:9.927e-05
## Max. :3.084954 Max. :2.247e-02library(ggplot2)
p <- plot(x = seq(1:20), y = gene_df[1:20,3]/217*100, ylab = 'Percentage of patients (%)', xlab = NA, main = 'Top mutated genes based on percentage of mutated patients', xaxt = "n") #xlab = 'Genes'
xtick <- seq(1:20)
axis(side = 1, at = xtick, labels = gene_df[1:20,1], las = 2, cex.axis = 0.75)
Save all the computed results in an excel file:
library('xlsx')
#write.xlsx(gene_df, file='KIRC_MUTATIONS.xlsx', sheetName = "mut",
# col.names = TRUE, row.names = FALSE, append = FALSE)