Clarify that a compound may have more annotated targets than indicated in the metadata

[shntnu]

The full list of targets is available here https://github.com/jump-cellpainting/jump-cellpainting/blob/master/0.design-pilots/output/drug_target_samples.csv

These 7 compounds are not listed in that CSV, so we'd need to fetch this from somewhere else, probably https://github.com/broadinstitute/lincs-cell-painting

broad_sample	gene	pert_id
BRD-K01824976-300-02-9	RPL23A	BRD-K01824976
BRD-K01825690-065-01-9	CYP1A2	BRD-K01825690
BRD-K75390981-300-04-9	HRH4	BRD-K75390981
BRD-K66430217-001-03-8	CLK1	BRD-K66430217
BRD-K01826724-003-07-9	KCTD16	BRD-K01826724
BRD-K01825589-310-02-9	KCNH7	BRD-K01825589
BRD-K01825660-322-01-9	TBXAS1	BRD-K01825660

[shntnu]

Notebook to produce

• the full target annotations: JUMP-Target_compound_metadata_all_targets.csv.zip
• connections across all ORFs, CRISPRs, and compounds after factoring in the full list of target JUMP-Target_compounds_crispr_orf_connections.csv.zip

R Notebook

library(tidyverse)

dfmc <- read_tsv("JUMP-Target-add_files/JUMP-Target_compound_metadata.tsv")

drug_target_samples <- read_csv("https://raw.githubusercontent.com/jump-cellpainting/jump-cellpainting/master/0.design-pilots/output/drug_target_samples.csv?token=AAJHQPEBFYWZ2CEVJUZ7IHC77BEAK")

dfmc_pert_iname <- 
  dfmc %>% 
  select(pert_iname) %>%
  mutate(pert_iname = str_split(pert_iname, "\\|")) %>%
  unnest(cols = c(pert_iname)) %>%
  mutate(pert_iname = str_trim(pert_iname)) %>%
  distinct()

dfmc_pert_iname %>%
  count() %>%
  knitr::kable()

n
304

dfmc_pert_iname %>%
  inner_join(drug_target_samples %>% distinct(pert_iname, target)) %>%
  distinct(pert_iname) %>%
  count() %>%
  knitr::kable()

## Joining, by = "pert_iname"

n
304

dfmc_full_targets <-
  dfmc %>%
  mutate(pert_iname = str_split(pert_iname, "\\|")) %>%
  unnest(cols = c(pert_iname)) %>%
  mutate(pert_iname = str_trim(pert_iname)) %>%
  distinct(pert_iname, broad_sample) %>%
  inner_join(dfmc_pert_iname %>%
               inner_join(drug_target_samples %>%
                            distinct(pert_iname, target)))

## Joining, by = "pert_iname"
## Joining, by = "pert_iname"

dfmc_full_targets <-
  bind_rows(dfmc_full_targets,
            dfmc %>%
              distinct(broad_sample, pert_iname, target)) %>%
  distinct(broad_sample, pert_iname, target)

dfmc_full_targets %>%
  distinct(pert_iname) %>%
  count() %>%
  knitr::kable()

n
305

dfmc_full_targets %>%
  mutate(pert_id = str_sub(broad_sample, 1, 13)) %>%
  distinct(pert_id) %>%
  count() %>%
  knitr::kable()

n
306

dfmc_full_targets %>%
  write_csv("JUMP-Target_compound_metadata_all_targets.csv")

dfmc <- read_tsv("JUMP-Target-add_files/JUMP-Target_compound_metadata.tsv")

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   broad_sample = col_character(),
##   InChIKey = col_character(),
##   pert_iname = col_character(),
##   pubchem_cid = col_double(),
##   target = col_character(),
##   pert_type = col_character(),
##   control_type = col_character(),
##   smiles = col_character()
## )

dfmo <- read_tsv("JUMP-Target-add_files/JUMP-Target_orf_metadata.tsv")

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   broad_sample = col_character(),
##   genes = col_character(),
##   pert_type = col_character(),
##   control_type = col_character()
## )

dfmx <- read_tsv("JUMP-Target-add_files/JUMP-Target_crispr_metadata.tsv")

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   broad_sample = col_character(),
##   genes = col_character(),
##   pert_type = col_character(),
##   control_type = col_character(),
##   target_sequence = col_character()
## )

dfmc1 <- dfmc_full_targets %>% distinct(broad_sample, pert_iname, target) %>% select(broad_sample_compound = broad_sample, pert_iname, gene = target)
dfmo1 <- dfmo %>% distinct(broad_sample, genes) %>% select(broad_sample_orf = broad_sample, gene = genes)
dfmx1 <- dfmx %>% distinct(broad_sample, genes) %>% select(broad_sample_crispr = broad_sample, gene = genes) %>% na.omit()

dfcg1 <- dfmc1 %>% distinct(gene) %>% na.omit() %>% pull("gene")
dfog <- dfmo1 %>% distinct(gene) %>% na.omit() %>% pull("gene")
dfxg <- dfmx1 %>% distinct(gene) %>% na.omit() %>% pull("gene")

connections <-
  dfmc1 %>% 
  inner_join(dfmo1) %>% 
  inner_join(dfmx1) %>%
  select(gene, broad_sample_compound, pert_iname, broad_sample_orf, broad_sample_crispr)

## Joining, by = "gene"

## Joining, by = "gene"

connections %>%
  distinct(gene) %>%
  count %>%
  knitr::kable()

n
160

connections %>%
  distinct(broad_sample_compound, broad_sample_orf) %>%
  count %>%
  knitr::kable()

n
466

connections %>%
  distinct(broad_sample_compound, broad_sample_crispr) %>%
  count %>%
  knitr::kable()

n
893

connections %>%
  distinct(broad_sample_compound, broad_sample_orf) %>%
  count %>%
  knitr::kable()

n
466

connections %>%
  distinct(broad_sample_orf, broad_sample_crispr) %>%
  count %>%
  knitr::kable()

n
305

connections %>%
  write_csv("JUMP-Target_compounds_crispr_orf_connections.csv")

connections %>%
  distinct(broad_sample_compound, broad_sample_orf) %>%
  group_by(broad_sample_orf) %>%
  tally(name = "n_compounds") %>% 
  group_by(n_compounds) %>%
  tally(name = "n_orfs") %>%
  select(n_orfs, n_compounds) %>% 
  knitr::kable()

n_orfs	n_compounds
6	1
83	2
34	3
19	4
8	5
4	6
1	7
2	8
1	9
2	10