Molecularly Subtyping Embryonal Tumors - Final Table

Libraries and functions

library(tidyverse)

Registered S3 method overwritten by 'dplyr':
  method           from
  print.rowwise_df     
── Attaching packages ─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── tidyverse 1.2.1 ──
✔ ggplot2 3.2.0     ✔ purrr   0.3.2
✔ tibble  2.1.3     ✔ dplyr   0.8.3
✔ tidyr   0.8.3     ✔ stringr 1.4.0
✔ readr   1.3.1     ✔ forcats 0.4.0
── Conflicts ────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()

# TODO: consider moving this out of this notebook to use with ependymoma 
# subtyping?
wrangle_fusions <- function(fusions_df, string_to_match, column_name) {
  # This function takes the fusion summary binary matrix and returns a 
  # data frame that contains two columns: Kids_First_Biospecimen_ID and
  # a column containing a comma-separated 'list' of relevant fusions.
  # A "relevant fusion" is determined by passing the string_to_match argument
  # to dplyr::matches. To specify that a gene symbol should be a 5' partner,
  # you can pass "<gene symbol>--". The 2nd column name is determined by the
  # column_name argument.
  #
  # Args:
  #   fusions_df: a fusion_summary data.frame from `fusion-summary`
  #   string_to_match: character; passed to dplyr::matches to select relevant
  #                    columns
  #   column_name: the name of the relevant fusion column in the data.frame 
  #                that is returned
  # 
  # Returns: a data.frame with Kids_First_Biospecimen_ID and relevant fusions
  #          columns; biospecimens without any relevant fusions will have "None"
  #          in the second column
  
  specific_fusions <- fusions_df %>%
    # only include relevant fusions using string_to_match
    select(Kids_First_Biospecimen_ID, matches(string_to_match)) %>%
    reshape2::melt() %>%
    # a value of 0 means that fusion is not present in that biospecimen
    filter(value > 0) %>%
    select(-value) %>%
    group_by(Kids_First_Biospecimen_ID) %>%
    # get the comma separated 'list' of relevant fusions
    summarize(!!column_name := paste(sort(unique(variable)), collapse = ", "))
  
  # if a sample had none of the relevant fusions, it's missing from 
  # specific_fusions -- let's add them back in with "None" in the second
  # column
  missing_ids <- setdiff(fusions_df$Kids_First_Biospecimen_ID,
                         specific_fusions$Kids_First_Biospecimen_ID)
  specific_fusions <- specific_fusions %>%
    bind_rows(data.frame(Kids_First_Biospecimen_ID = missing_ids)) %>%
    replace(is.na(.), "None")
  
  return(specific_fusions)
}

Directories

subset_dir <- "subset-files"
results_dir <- "results"
data_dir <- file.path("..", "..", "data")

Read in files

Set up filenames.

# full clinical file
histologies_file <- file.path(data_dir, "pbta-histologies.tsv")
# TODO: update when the new files with union of sample IDs get included in the
# in the data download (v14)
fusion_file <- file.path("..", "fusion-summary", "results",
                         "fusion_summary_embryonal_foi.tsv")
manta_file <- file.path(subset_dir, "embryonal_manta_sv.tsv")
polya_file <- file.path(subset_dir, "embryonal_zscored_exp.polya.rds")
stranded_file <- file.path(subset_dir, "embryonal_zscored_exp.stranded.rds")
# file that contains the biospecimen IDs that met criteria for subtyping
biospecimen_file <- file.path(results_dir, 
                              "biospecimen_ids_embryonal_subtyping.tsv")
# cleaned chr19 data
chr19_file <- file.path(results_dir, "cleaned_chr19_cn.tsv")
# annotated sex chromosomes copy number alterations
sex_chr_consensus <- file.path("..", "focal-cn-file-preparation", "results", 
                               "consensus_seg_annotated_cn_x_and_y.tsv.gz")
sex_chr_cnvkit <- file.path("..", "focal-cn-file-preparation", "results", 
                            "cnvkit_annotated_cn_x_and_y.tsv.gz")
sex_chr_controlfreec <- file.path("..", "focal-cn-file-preparation", "results", 
                                  "controlfreec_annotated_cn_x_and_y.tsv.gz")

Read in files.

histologies_df <- read_tsv(histologies_file, 
                           col_types = cols(
                             molecular_subtype = col_character()
                          ))

# manta data for BCOR tandem duplication
manta_sv_df <- read_tsv(manta_file)

# all of the RNA data
polya_exp <- read_rds(polya_file)
stranded_exp <- read_rds(stranded_file)
fusions_df <- read_tsv(fusion_file)

# get biospecimen ids as a vector rather than data frame
biospecimen_ids <- read_tsv(biospecimen_file) %>%
  pull(Kids_First_Biospecimen_ID)

# cleaned chr19 amplification data
chr19_cn_df <- read_tsv(chr19_file)

# we'll filter to only the relevant biospecimens for the copy number alteration
# data that is committed to the repository
consensus_sex_chr <- read_tsv(sex_chr_consensus) %>%
  filter(biospecimen_id %in% biospecimen_ids)
cnvkit_sex_chr <- read_tsv(sex_chr_cnvkit) %>%
  filter(biospecimen_id %in% biospecimen_ids)
controlfreec_sex_chr <- read_tsv(sex_chr_controlfreec) %>%
  filter(biospecimen_id %in% biospecimen_ids)

Output file

output_all <- file.path(results_dir, 
                        "embryonal_tumor_subtyping_relevant_data.tsv")
output_subtypes  <- file.path(results_dir, 
                              "embryonal_tumor_molecular_subtypes.tsv")

Fusion data

Summarizing the salient points from AlexsLemonade/OpenPBTA-analysis#251:

ETMR, C19MC-altered: These tumors have focal amplification of the miRNA cluster on chr19 (denoted C19MC) and often have gene fusions involving TTYH1 and chr19 miRNA cluster genes

CNS HGNET-MN1: Contain gene fusions involving 5’ MN1. 3’ fusion partners can include BEND2 and CXXC5.

CNS NB-FOXR2: Over-expression and/or gene fusions in FOXR2

CNS EFT-CIC: Alterations in CIC, commonly fused with NUTM1

colnames(fusions_df)

 [1] "Kids_First_Biospecimen_ID" "MACF1--TTYH1"              "MN1--AC008667.1"           "MN1--CXXC5"                "MN1--PATZ1"                "MN1--PSD2"                
 [7] "NPAS3--AC245884.4/TTYH1"   "TTYH1--DPRX/RNU6-698P"     "TTYH1--MIR512-2/MIR1323"   "TTYH1--MIR515-2/MIR519C"   "ZC4H2--AL627224.1/FOXR2"   "ZC4H2--FOXR2"             
[13] "CIC--NUTM1"                "MN1--BEND2"               

ttyh1_fusions <- wrangle_fusions(fusions_df, "TTYH1--", "TTYH1_fusions")

binding character and factor vector, coercing into character vector

mn1_fusions <- wrangle_fusions(fusions_df, "MN1--", "MN1_fusions") 

binding character and factor vector, coercing into character vector

foxr2_fusions <- wrangle_fusions(fusions_df, "FOXR2", "FOXR2_fusions")

binding character and factor vector, coercing into character vector

cic_fusions <- wrangle_fusions(fusions_df, "CIC--", "CIC_fusions")

binding character and factor vector, coercing into character vector

fusions_summary_df <- list(ttyh1_fusions,
                           mn1_fusions,
                           foxr2_fusions,
                           cic_fusions) %>%
  purrr::reduce(dplyr::inner_join, by = "Kids_First_Biospecimen_ID")
# remove data.frame we no longer need
rm(ttyh1_fusions, mn1_fusions, foxr2_fusions, cic_fusions, fusions_df)

head(fusions_summary_df %>% arrange(desc(TTYH1_fusions)), n = 10)

Expression data

Summarizing salient points from AlexsLemonade/OpenPBTA-analysis#251:

ETMR, C19MC-altered: These tumors have high expression of LIN28A and serves as a biomarker for ETMRs

CNS NB-FOXR2: Over-expression and/or gene fusions in FOXR2

ETMR, NOS: These tumors have high expression of LIN28A and serves as a biomarker for ETMRs, but do not show focal amplification of C19MC.

exp_genes_of_interest <- c("LIN28A", "FOXR2")
# set up poly-A data
polya_exp <- polya_exp[, exp_genes_of_interest] %>%
  as.data.frame() %>%
  tibble::rownames_to_column("Kids_First_Biospecimen_ID")
# set up stranded data
stranded_exp <- stranded_exp[, exp_genes_of_interest] %>%
  as.data.frame() %>%
  tibble::rownames_to_column("Kids_First_Biospecimen_ID")

# bind together both datasets, but keep track of what selection strategy was
# used
exp_df <- bind_rows("polya" = polya_exp, 
                    "stranded" = stranded_exp, 
                    .id = "exp_dataset") %>%
  select(-exp_dataset, everything())

# remove data.frame we no longer need
rm(polya_exp, stranded_exp, exp_genes_of_interest)

Join RNA data

rna_df <- exp_df %>%
  left_join(fusions_summary_df) %>%
  select(-exp_dataset, everything())

Joining, by = "Kids_First_Biospecimen_ID"

# remove data.frame we no longer need
rm(exp_df)

head(rna_df, n = 10)

We see some repeated negative values for the expression z-scores here. Generally that comes from small or zero log2(FPKM + 1) values. For example, all of the poly-A values for FOXR2 are zero prior to scaling.

Structural variant data

CNS HGNET-BCOR

• CNS high-grade neuroepithelial tumor with BCOR alteration

• Tumors have internal tandem duplication of BCOR

The definition of internal tandem duplication from Rustagi et al. BMC Bioinformatics. 2016.:

Detection of tandem duplication within coding exons, referred to as internal tandem duplication (ITD)

The annotated files from the focal-cn-file-preparation module use exons to check for overlaps. So, we can try to supplement the Manta calls with the annotated sex chromosome files.

# only include rows from Manta that make reference to BCOR, pass all filters
# https://github.com/Illumina/manta/blob/75b5c38d4fcd2f6961197b28a41eb61856f2d976/docs/userGuide/README.md#vcf-format-fields
# and are recorded as tandem duplication events
bcor_sv_df <- manta_sv_df %>%
  filter(str_detect(Gene.name, "BCOR"),
         FILTER == "PASS",
         ALT == "<DUP:TANDEM>")

# take a look at the genes 
bcor_sv_df %>% 
  head(n = 2) %>%
  pull(Gene.name)

[1] "ABCB7/ACE2/ACOT9/ADGRG2/AKAP4/ALAS2/AMELX/AMER1/ANOS1/AP1S2/APEX2/APOO/APOOL/AR/ARAF/ARHGAP6/ARHGEF9/ARHGEF9-IT1/ARR3/ARX/ASB11/ASB12/ASB9/ATP6AP2/ATP7A/ATRX/ATXN3L/AWAT1/AWAT2/BCLAF3/BCOR/BEND2/BMP15/BMX/BRDTP1/BRWD3/CA5B/CA5BP1/CA5BP1-CA5B/CACNA1F/CASK/CBLL2/CCDC120/CCDC22/CCNB3/CDK16/CDKL5/CDX4/CENPVL1/CENPVL2/CENPVL3/CFAP47/CFP/CHIC1/CHM/CHMP1B2P/CHST7/CITED1/CLCN4/CLCN5/CLDN34/CLTRN/CNKSR2/COX7B/CPXCR1/CTPS2/CXCR3/CXXC1P1/CXorf21/CXorf38/CXorf49/CXorf49B/CXorf58/CXorf65/CXorf67/CYBB/CYLC1/CYSLTR1/DACH2/DCAF8L1/DCAF8L2/DDX3X/DDX53/DGAT2L6/DGKK/DIAPH2/DIAPH2-AS1/DIPK2B/DLG3/DLG3-AS1/DMD/DMRTC1/DMRTC1B/DUSP21/DYNLT3/EBP/EDA/EDA2R/EFHC2/EFNB1/EGFL6/EIF1AX/EIF1AX-AS1/EIF2S3/ELK1/ERAS/ERCC6L/FAAH2/FAM104B/FAM120C/FAM133A/FAM155B/FAM156A/FAM156B/FAM226A/FAM226B/FAM236A/FAM236B/FAM236C/FAM236D/FAM47A/FAM47B/FAM47C/FAM9A/FAM9B/FAM9C/FANCB/FGD1/FGF16/FLICR/FLJ44635/FOXO4/FOXP3/FOXR2/FRMD8P1/FRMPD4/FTH1P18/FTHL17/FTSJ1/FTX/FUNDC1/GAGE1/GAGE10/GAGE12B/GAGE12C/GAGE12D/GAGE12E/GAGE12F/GAGE12G/GAGE12H/GAGE12I/GAGE12J/GAGE13/GAGE2A/GAGE2B/GAGE2C/GAGE2D/GAGE2E/GAGE4/GAGE5/GAGE6/GAGE7/GAGE8/GATA1/GCNA/GDPD2/GEMIN8/GJB1/GK/GLOD5/GLRA2/GNL3L/GPKOW/GPM6B/GPR143/GPR173/GPR174/GPR34/GPR82/GRIPAP1/GRPR/GS1-594A7.3/GS1-600G8.3/GSPT2/HCCS/HDAC6/HDAC8/HDX/HEPH/HMGN5/HSD17B10/HUWE1/HYPM/IGBP1/IL1RAPL1/IL2RG/INE1/INE2/INGX/IQSEC2/ITGB1BP2/ITIH6/ITM2A/JADE3/JPX/KANTR/KCND1/KDM5C/KDM6A/KIF4A/KLF8/KLHL15/KLHL34/KLHL4/KRBOX4/LANCL3/LAS1L/LINC00269/LINC00891/LINC01186/LINC01203/LINC01204/LINC01278/LINC01281/LINC01282/LINC01284/LINC01456/LINC01496/LINC01545/LINC01560/LINC02154/LINC02595/LINC02601/LOC100129291/LOC100132741/LOC100132831/LOC100421746/LOC101059915/LOC101927476/LOC101927635/LOC101928128/LOC101928201/LOC101928389/LOC101928627/LOC105373156/LOC392452/LOC729609/LPAR4/MAGEB1/MAGEB10/MAGEB16/MAGEB17/MAGEB18/MAGEB2/MAGEB3/MAGEB4/MAGEB5/MAGEB6/MAGED1/MAGED2/MAGED4/MAGED4B/MAGEE1/MAGEE2/MAGEH1/MAGIX/MAGT1/MAOA/MAOB/MAP2K4P1/MAP3K15/MAP7D2/MBTPS2/MED12/MED14/MED14OS/MID1/MID1IP1/MID1IP1-AS1/MIR1321/MIR1468/MIR1587/MIR188/MIR221/MIR222/MIR223/MIR23C/MIR325/MIR325HG/MIR361/MIR362/MIR374A/MIR374B/MIR374C/MIR384/MIR3915/MIR3937/MIR421/MIR4328/MIR4454/MIR4536-1/MIR4536-2/MIR4666B/MIR4767/MIR4768/MIR4769/MIR4770/MIR500A/MIR500B/MIR501/MIR502/MIR532/MIR545/MIR548AJ2/MIR548AM/MIR548AX/MIR548F5/MIR548I4/MIR548M/MIR6086/MIR6134/MIR651/MIR660/MIR676/MIR6857/MIR6894/MIR6895/MIR8088/MIR98/MIRLET7F2/MOSPD2/MPC1L/MSL3/MSN/MTMR8/MTRNR2L10/NAP1L2/NAP1L3/NAP1L6/NBDY/NDP/NDP-AS1/NDUFB11/NEXMIF/NHS/NHS-AS1/NHSL2/NLGN3/NLGN4X/NLRP2B/NONO/NR0B1/NUDT10/NUDT11/NYX/OFD1/OGT/OPHN1/OTC/OTUD5/OTUD6A/P2RY10/P2RY4/PABPC1L2A/PABPC1L2B/PABPC1L2B-AS1/PABPC5/PABPC5-AS1/PAGE1/PAGE2/PAGE2B/PAGE3/PAGE4/PAGE5/PBDC1/PCDH11X/PCSK1N/PCYT1B/PCYT1B-AS1/PDHA1/PDK3/PDZD11/PFKFB1/PGAM4/PGK1/PHEX/PHEX-AS1/PHF8/PHKA1/PHKA1-AS1/PHKA2/PHKA2-AS1/PIGA/PIM2/PIN4/PINCR/PIR/PIR-FIGF/PJA1/PLP2/PNPLA4/POF1B/POLA1/PORCN/POU3F4/PPEF1/PPEF1-AS1/PPP1R2C/PPP1R3F/PPP4R3C/PQBP1/PRAF2/PRDX4/PRICKLE3/PRPS2/PRRG1/PTCHD1/PTCHD1-AS/PUDP/RAB41/RAB9A/RAI2/RBBP7/RBM10/RBM3/REPS2/RGN/RIBC1/RLIM/RP2/RPA4/RPGR/RPS26P11/RPS4X/RPS6KA3/RPS6KA6/RRAGB/RS1/RTL3/RTL5/S100G/SAT1/SATL1/SCARNA23/SCARNA9L/SCML1/SCML2/SH3BGRL/SH3KBP1/SHROOM2/SHROOM4/SLC16A2/SLC35A2/SLC38A5/SLC7A3/SLC9A7/SMC1A/SMPX/SMS/SNORA109/SNORA11/SNORA11C/SNORA11D/SNORA11E/SNORA11G/SNORD3E/SNX12/SPACA5/SPACA5B/SPANXN5/SPIN2A/SPIN2B/SPIN3/SPIN4/SRPX/SSX1/SSX2/SSX2B/SSX3/SSX4/SSX4B/SSX5/SSX6P/SSX7/SSX8P/SSX9P/STARD8/STS/SUPT20HL1/SUPT20HL2/SUV39H1/SYAP1/SYN1/SYP/SYP-AS1/SYTL5/TAB3/TAB3-AS1/TAF1/TAF9B/TBC1D25/TBL1X/TBX22/TCEANC/TENT5D/TEX11/TFE3/TGIF2LX/TIMM17B/TIMP1/TLR7/TLR8/TLR8-AS1/TMEM47/TMSB4X/TRAPPC2/TRO/TSIX/TSPAN7/TSPYL2/TSR2/TTC3P1/TXLNG/UBA1/UBE2DNL/UBE2E4P/UBQLN2/UPRT/UQCRBP1/USP11/USP27X/USP27X-AS1/USP51/USP9X/UXT/UXT-AS1/VCX/VCX2/VCX3A/VCX3B/VEGFD/VENTXP1/VSIG4/WAS/WDR13/WDR45/WNK3/WWC3/XAGE1A/XAGE1B/XAGE2/XAGE3/XAGE5/XIST/XK/YIPF6/YY2/ZC3H12B/ZC4H2/ZCCHC13/ZDHHC15/ZFX/ZFX-AS1/ZMYM3/ZNF157/ZNF182/ZNF41/ZNF630/ZNF630-AS1/ZNF674/ZNF674-AS1/ZNF711/ZNF81/ZRSR2/ZXDA/ZXDB"
[2] "ACE2/ACOT9/ADGRG2/AMELX/ANOS1/AP1S2/APOO/ARHGAP6/ARX/ASB11/ASB9/ATP6AP2/ATXN3L/BCLAF3/BCOR/BEND2/BMX/CA5B/CA5BP1/CA5BP1-CA5B/CASK/CBLL2/CDKL5/CFAP47/CLCN4/CLDN34/CLTRN/CNKSR2/CTPS2/CXorf21/CXorf38/CXorf58/CYBB/DCAF8L1/DCAF8L2/DDX3X/DDX53/DIPK2B/DMD/DUSP21/DYNLT3/EFHC2/EGFL6/EIF1AX/EIF1AX-AS1/EIF2S3/FAM47A/FAM47B/FAM47C/FAM9A/FAM9B/FAM9C/FANCB/FRMPD4/FTH1P18/FTHL17/FUNDC1/GEMIN8/GK/GLRA2/GPM6B/GPR143/GPR34/GPR82/GRPR/GS1-594A7.3/GS1-600G8.3/HCCS/HYPM/IL1RAPL1/INE2/KDM6A/KLHL15/KLHL34/KRBOX4/LANCL3/LINC01186/LINC01203/LINC01204/LINC01281/LINC01282/LINC01456/LINC02154/LINC02595/LINC02601/LOC100132831/LOC101927476/LOC101928389/LOC101928627/LOC392452/LOC729609/MAGEB1/MAGEB10/MAGEB16/MAGEB17/MAGEB18/MAGEB2/MAGEB3/MAGEB4/MAGEB5/MAGEB6/MAOA/MAOB/MAP3K15/MAP7D2/MBTPS2/MED14/MED14OS/MID1/MID1IP1/MID1IP1-AS1/MIR1587/MIR221/MIR222/MIR23C/MIR3915/MIR3937/MIR4666B/MIR4767/MIR4768/MIR548AJ2/MIR548AM/MIR548AX/MIR548F5/MIR6086/MIR6134/MIR651/MOSPD2/MPC1L/MSL3/NDP/NDP-AS1/NHS/NHS-AS1/NR0B1/NYX/OFD1/OTC/PCYT1B/PCYT1B-AS1/PDHA1/PDK3/PHEX/PHEX-AS1/PHKA2/PHKA2-AS1/PIGA/PINCR/PIR/PIR-FIGF/PNPLA4/POLA1/PPEF1/PPEF1-AS1/PPP1R2C/PPP4R3C/PRDX4/PRPS2/PRRG1/PTCHD1/PTCHD1-AS/PUDP/RAB9A/RAI2/RBBP7/REPS2/RPGR/RPS6KA3/RS1/S100G/SAT1/SCARNA23/SCARNA9L/SCML1/SCML2/SH3KBP1/SHROOM2/SMPX/SMS/SRPX/STS/SUPT20HL1/SUPT20HL2/SYAP1/SYTL5/TAB3/TAB3-AS1/TBL1X/TCEANC/TLR7/TLR8/TLR8-AS1/TMEM47/TMSB4X/TRAPPC2/TSPAN7/TXLNG/UBE2E4P/USP9X/VCX/VCX2/VCX3B/VEGFD/VENTXP1/WWC3/XK/YY2/ZFX/ZFX-AS1/ZNF674/ZNF674-AS1/ZRSR2"                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           

Not necessarily specific to BCOR alterations.

consensus_sex_chr %>% filter(gene_symbol == "BCOR") 

cnvkit_sex_chr %>% filter(gene_symbol == "BCOR") 

controlfreec_sex_chr %>% filter(gene_symbol == "BCOR")

This is not conclusive.

# remove all the sv and cnv data we will not use
rm(bcor_sv_df, manta_sv_df, consensus_sex_chr, controlfreec_sex_chr,
   cnvkit_sex_chr)

Add Kids_First_Participant_ID and sample_id

identifiers_df <- histologies_df %>%
  select(Kids_First_Participant_ID, 
         sample_id, 
         Kids_First_Biospecimen_ID)
# add IDs to chromosome 19 data
chr19_cn_df <- identifiers_df %>%
  inner_join(chr19_cn_df, by = "Kids_First_Biospecimen_ID")
# add IDs to RNA data 
rna_df <- identifiers_df %>%
  inner_join(rna_df)

Joining, by = "Kids_First_Biospecimen_ID"

Join molecular data

molecular_data_df <- full_join(chr19_cn_df, rna_df,
                               by = c("Kids_First_Participant_ID",
                                      "sample_id")) %>%
  rename(Kids_First_Biospecimen_ID_DNA = Kids_First_Biospecimen_ID.x,
         Kids_First_Biospecimen_ID_RNA = Kids_First_Biospecimen_ID.y)

Add clinical data

The following were mentioned in AlexsLemonade/OpenPBTA-analysis#251:

CNS HGNET-BCOR: Median age of diagnosis less than 10 years

CNS HGNET-MN1: Predominantly female patients.

CNS Embryonal, NOS: Tumors previously called PNET that do not fit into other groups above.

relevant_clinical_data <- histologies_df %>% 
  filter(Kids_First_Biospecimen_ID %in% biospecimen_ids) %>%
  select(Kids_First_Participant_ID, 
         sample_id,
         age_at_diagnosis_days,
         germline_sex_estimate,
         disease_type_old,
         disease_type_new) %>%
  distinct() %>%
  # convert age to years
  mutate(age_at_diagnosis_yrs = 
           floor(as.numeric(age_at_diagnosis_days) / 365)) %>%
  select(-age_at_diagnosis_days)

all_data_df <- inner_join(relevant_clinical_data,
                          molecular_data_df)

Joining, by = c("Kids_First_Participant_ID", "sample_id")

Write to file.

all_data_df %>%
  select(Kids_First_Participant_ID, 
         sample_id,
         Kids_First_Biospecimen_ID_DNA,
         Kids_First_Biospecimen_ID_RNA,
         age_at_diagnosis_yrs,
         germline_sex_estimate,
         disease_type_old,
         disease_type_new,
         LIN28A,
         TTYH1_fusions,
         chr19_amplification,
         FOXR2,
         FOXR2_fusions,
         MN1_fusions,
         CIC_fusions) %>%
  write_tsv(output_all)

ETMR, C19MC-altered and ETMR, NOS

The differences between ETMR, C19MC-altered and ETMR, NOS is a difference in C19MC amplification. We’ll filter on LIN28A overexpression, as that is a feature of both.

Recall that we z-scored these expression values before subsetting the matrix so they are in the context of all samples for a given selection strategy.

all_data_df %>%
  # 3 standard deviations above the mean for LIN28A expression or
  # evidence of C19MC amplification
  filter(LIN28A > 3 | chr19_amplification == "Yes") %>%
  select(ends_with("ID"), 
         starts_with("disease"),
         LIN28A,
         TTYH1_fusions,
         chr19_amplification)

For some samples that have a TTYH1 fusion, we do not have DNA data to check for C19MC amplification but that’s okay per this comment which references Kleinman et al. Nat Genet. 2014.. We will call all samples with a TTYH1 fusion ETMR, C19MC altered.

etmr_c19mc_df <- all_data_df %>%
  filter(LIN28A > 3 | chr19_amplification == "Yes",
         TTYH1_fusions != "None") %>%
  select(contains("ID")) %>%
  mutate(molecular_subtype = "ETMR, C19MC-altered")

For now, we will call the sample with a LIN28A z-score > 3 that has no TTYH1 fusion and no copy number data ETMR, NOS.

# TODO: is this the correct call or should this be a CNS Embryonal, NOS call?
etmr_nos_df <- all_data_df %>%
  filter(LIN28A > 3,
         TTYH1_fusions == "None") %>%
  select(contains("ID")) %>%
  mutate(molecular_subtype = "ETMR, NOS")

CNS HGNET-MN1

• CNS high-grade neuroepithelial tumor with MN1 alteration
• Likely previously diagnosed as PNET.
• Contain gene fusions involving 5’ MN1. 3’ fusion partners can include BEND2 and CXXC5.
• Predominantly female patients.

all_data_df %>%
  # 3 standard deviations above the mean
  filter(MN1_fusions != "None") %>%
  select(ends_with("ID"), 
         starts_with("disease"),
         germline_sex_estimate,
         MN1_fusions)

No samples fit this criteria. There were other samples that did contain MN1 fusions.

mn1_fusions <- fusions_summary_df %>%
  filter(MN1_fusions != "None") 

mn1_fusions

histologies_df %>%
  filter(Kids_First_Biospecimen_ID %in% pull(mn1_fusions, 
                                             Kids_First_Biospecimen_ID)) %>%
  select(Kids_First_Participant_ID, 
         sample_id, 
         Kids_First_Biospecimen_ID,
         disease_type_old,
         disease_type_new,
         germline_sex_estimate)

They are classified as Ependymoma and Low-grade astrocytic tumor.

We will call the sample with the MN1–CXXC5 fusion CNS HGNET-MN1.

hgnet_mn1_df <- histologies_df %>%
  filter(sample_id == "7316-765") %>%
  select(Kids_First_Participant_ID,
         Kids_First_Biospecimen_ID_RNA = Kids_First_Biospecimen_ID,
         sample_id) %>%
  mutate(molecular_subtype = "CNS HGNET-MN1")

CNS HGNET-BCOR

• CNS high-grade neuroepithelial tumor with BCOR alteration
• Tumors have internal tandem duplication of BCOR
• Median age of diagnosis less than 10 years

We have no clear evidence of BCOR alterations.

CNS NB-FOXR2

• Central nervous system (CNS) neuroblastoma with FOXR2 activation
• Over-expression and/or gene fusions in FOXR2

all_data_df %>%
  filter(FOXR2_fusions != "None" | FOXR2 > 3) %>%
  select(ends_with("ID"), 
         starts_with("disease"),
         FOXR2,
         FOXR2_fusions)

nb_foxr2_df <- all_data_df %>%
  filter(FOXR2_fusions != "None" | FOXR2 > 3) %>%
  select(contains("ID")) %>%
  mutate(molecular_subtype = "CNS NB-FOXR2")

CNS EFT-CIC

• CNS Ewing sarcoma family tumor with CIC alteration
• Alterations in CIC, commonly fused with NUTM1

all_data_df %>%
  filter(CIC_fusions != "None")

CNS Embryonal, NOS

• CNS Embryonal tumor, not otherwise specified
• Tumors previously called PNET that do not fit into other groups above.

All other samples will be called CNS Embryonal, NOS.

Subtyping file

subtypes_df <- bind_rows(
  etmr_c19mc_df,
  etmr_nos_df,
  hgnet_mn1_df,
  nb_foxr2_df
)

subtypes_df <- all_data_df %>%
  select(contains("ID")) %>%
  left_join(subtypes_df) %>%
  replace_na(list(molecular_subtype = "CNS Embryonal, NOS"))

Joining, by = c("Kids_First_Participant_ID", "sample_id", "Kids_First_Biospecimen_ID_DNA", "Kids_First_Biospecimen_ID_RNA")

subtypes_df %>%
  write_tsv(output_subtypes)