Parse AntiSmash and BigScape

This notebook parses AntiSmash and BigScape output and it is heavily inspired in the work of Roberto Sánchez.

https://github.com/robertosanchezn/AS_hqMAGs/blob/main/r_markdown/notebook/parse_bigscape.R

Antismash

Set up

library(RauENVS)
antismash_directory <- system.file(
  "antismash_example_directory",package =  "RauENVS"
  )
# Edit this to your actual file!!! and remove or comment the example file
# For example:
# antismash_directory <- "/mnt"

First, we need to identify the json files:

list_jsons <- function(dir) list.files(
  dir, ".json$", full.names = T,recursive = T,include.dirs = T
  )
(jsons <- list_jsons(antismash_directory))
#> [1] "/home/runner/work/_temp/Library/RauENVS/antismash_example_directory/antismash.json"
(genome_ids <- stringr::str_match(basename(jsons), "(^.+)\\.json")[,2])
#> [1] "antismash"
names(jsons) <- genome_ids

Get features

Now, we want to extract the features for each json file.

features <- get_antismash_features(jsons, c("region", "aSModule"))

Features is a large list, so inspect it carefully. It should have as many lists as json files:

stopifnot(length(features) == length(jsons))

Regions

Now, we are extracting all regions:

regions <- get_antismash_regions(features)
regions |>
  head(30) |>
  knitr::kable()

file	location	type	contig_edge	product	bgc_id
antismash	[292344:342785]	region	FALSE	bacteriocin, NRPS	.region001
antismash	[999349:1105105]	region	FALSE	transAT-PKS , transAT-PKS-like	.region002
antismash	[1222307:1263002]	region	FALSE	T3PKS	.region003
antismash	[1347426:1367554]	region	FALSE	terpene	.region004
antismash	[1397403:1532936]	region	FALSE	betalactone, NRPS , transAT-PKS	.region005
antismash	[1724843:1834411]	region	FALSE	NRPS , T3PKS , transAT-PKS , transAT-PKS-like	.region006
antismash	[2053604:2141836]	region	FALSE	transAT-PKS	.region007
antismash	[2448155:2468896]	region	FALSE	terpene	.region008
antismash	[2553885:2595130]	region	FALSE	PKS-like	.region009
antismash	[2824697:2865819]	region	FALSE	ladderane	.region010
antismash	[3163092:3228500]	region	FALSE	NRPS	.region011
antismash	[3855452:3896871]	region	FALSE	other	.region012

You will notice that the product column is nested (each cell is a list of values). You can either paste the different values:

regions |>
  dplyr::mutate(
    product = purrr::map(product, paste0, collapse = ",") |>
      as.character()
    ) |>
  head(30) |>
  knitr::kable()

file	location	type	contig_edge	product	bgc_id
antismash	[292344:342785]	region	FALSE	bacteriocin,NRPS	.region001
antismash	[999349:1105105]	region	FALSE	transAT-PKS,transAT-PKS-like	.region002
antismash	[1222307:1263002]	region	FALSE	T3PKS	.region003
antismash	[1347426:1367554]	region	FALSE	terpene	.region004
antismash	[1397403:1532936]	region	FALSE	betalactone,NRPS,transAT-PKS	.region005
antismash	[1724843:1834411]	region	FALSE	NRPS,T3PKS,transAT-PKS,transAT-PKS-like	.region006
antismash	[2053604:2141836]	region	FALSE	transAT-PKS	.region007
antismash	[2448155:2468896]	region	FALSE	terpene	.region008
antismash	[2553885:2595130]	region	FALSE	PKS-like	.region009
antismash	[2824697:2865819]	region	FALSE	ladderane	.region010
antismash	[3163092:3228500]	region	FALSE	NRPS	.region011
antismash	[3855452:3896871]	region	FALSE	other	.region012

Or you can unnest the DataFrame by creating a new row per each product:

regions |>
  tidyr::unnest("product")|>
  head(30) |>
  knitr::kable()

file	location	type	contig_edge	product	bgc_id
antismash	[292344:342785]	region	FALSE	bacteriocin	.region001
antismash	[292344:342785]	region	FALSE	NRPS	.region001
antismash	[999349:1105105]	region	FALSE	transAT-PKS	.region002
antismash	[999349:1105105]	region	FALSE	transAT-PKS-like	.region002
antismash	[1222307:1263002]	region	FALSE	T3PKS	.region003
antismash	[1347426:1367554]	region	FALSE	terpene	.region004
antismash	[1397403:1532936]	region	FALSE	betalactone	.region005
antismash	[1397403:1532936]	region	FALSE	NRPS	.region005
antismash	[1397403:1532936]	region	FALSE	transAT-PKS	.region005
antismash	[1724843:1834411]	region	FALSE	NRPS	.region006
antismash	[1724843:1834411]	region	FALSE	T3PKS	.region006
antismash	[1724843:1834411]	region	FALSE	transAT-PKS	.region006
antismash	[1724843:1834411]	region	FALSE	transAT-PKS-like	.region006
antismash	[2053604:2141836]	region	FALSE	transAT-PKS	.region007
antismash	[2448155:2468896]	region	FALSE	terpene	.region008
antismash	[2553885:2595130]	region	FALSE	PKS-like	.region009
antismash	[2824697:2865819]	region	FALSE	ladderane	.region010
antismash	[3163092:3228500]	region	FALSE	NRPS	.region011
antismash	[3855452:3896871]	region	FALSE	other	.region012

Remember to “save” your processed dataframe before saving it into a file. For example:

regions_unnest <- regions |>
  tidyr::unnest("product")
#write.csv(regions_unnest, "regions.csv")

Modules

Now, we extract the modules:

modules <- get_antismash_modules(features)
# write.csv(modules, "/mnt/modules.csv")
modules |>
  head(30) |>
  knitr::kable()

file	location	complete	type	iterative
antismash	[295017:295515]	FALSE	unknown	FALSE
antismash	[312427:312892]	FALSE	unknown	FALSE
antismash	[314517:315729]	FALSE	nrps	FALSE
antismash	[316751:316937]	FALSE	unknown	FALSE
antismash	[317020:320092]	TRUE	nrps	FALSE
antismash	[320134:324067]	TRUE	nrps	FALSE
antismash	[1019210:1020041]	FALSE	pks	FALSE
antismash	[1022529:1022739]	FALSE	unknown	FALSE
antismash	[1022816:1023860]	FALSE	nrps	FALSE
antismash	[1024165:1024666]	FALSE	unknown	FALSE
antismash	[1024939:1025149]	FALSE	unknown	FALSE
antismash	[1025281:1030747]	TRUE	pks	FALSE
antismash	[1030891:1031074]	FALSE	unknown	FALSE
antismash	[1031221:1035418]	TRUE	pks	FALSE
antismash	[1035550:1037368]	FALSE	pks	FALSE
antismash	[1037719:1038733]	FALSE	unknown	FALSE
antismash	[1038892:1042354]	TRUE	pks	FALSE
antismash	[1042477:1043683]	FALSE	pks	FALSE
antismash	[1043898:1045017]	FALSE	unknown	FALSE
antismash	[1045152:1047294]	TRUE	pks	FALSE
antismash	[1047402:1049409]	FALSE	unknown	FALSE
antismash	[1050048:1051038]	FALSE	unknown	FALSE
antismash	[1051128:1051326]	FALSE	unknown	FALSE
antismash	[1051509:1055778]	TRUE	pks	FALSE
antismash	[1055943:1060086]	TRUE	pks	FALSE
antismash	[1060239:1063737]	TRUE	pks	FALSE
antismash	[1063857:1065108]	FALSE	pks	FALSE
antismash	[1065322:1066471]	FALSE	unknown	FALSE
antismash	[1067089:1067635]	FALSE	unknown	FALSE
antismash	[1067905:1071412]	TRUE	pks	FALSE

BigScape

bigscape_directory <- system.file(
  "bigscape_example_directory",package =  "RauENVS"
  )
# Edit this to your actual file!!! and remove or comment the example file
# For example:
# bigscape_directory <- "/mnt/"

First, we will create a DataFrame with clustering information:

bigscape_clusters <- get_bigscape_clustering(bigscape_directory)
#write_csv(bigscape_clusters, "/mnt/bigscape_df.csv")
bigscape_clusters |>
  head(30) |>
  knitr::kable()

bgc_id	class	GCF_c0.30
10B3s2	Others	12
10B4	Others	16
5B2new	Others	80
5C1	Others	92
5C1_4.region001	Others	92
CCC2s4new	Others	92
CCC2s4new_14.region001	Others	92
10B2	Others	95
10B2new	Others	95
10I4	Others	95
5B2	Others	95
5C4	Others	95
10C2	Others	125
10C2_4.region001	Others	125
10I1	Others	125
10I1_13.region001	Others	125
10I1new_43.region001	Others	125
10S3	Others	125
10S3_14.region001	Others	125
5B3	Others	125
5I3	Others	125
5I3_1.region001	Others	125
5I3new	Others	125
5I3new_13.region001	Others	125
RTC2s1	Others	125
RTC2s1_3.region001	Others	125
RTC2s1new	Others	125
RTS3_13.region001	Others	125
5I2	Others	190
5I2_2.region001	Others	190

We can extract more information. For example, for finding GCFs

library(tidyverse)
bigscape_clusters |>
    select(bgc_id, starts_with('GCF')) |>
    filter(str_detect(bgc_id, "^BGC\\d{7}\\.1")) |>
    select(-bgc_id) |>
    pull()
#> character(0)

Finally, we read network information:

bigscape_networks <- get_bigscape_networks(bigscape_directory)
#write_csv(bigscape_networks, "/mnt/bigscape_networks.csv")
bigscape_networks |>
  head(30) |>
  knitr::kable()