正則表達式提取 R 中字符之間的字符串

Question

我有一個非常混亂的文件，有多個分隔符，並且行之間的字段順序/數量不同。

V1 到 V5 列很好，但我想從 V9 中提取來自“Variant_seq”、“Reference_seq”的信息和來自“Dbxref”的 rsxxxx 編號。

另一個復雜之處是“Variant_seq”和“Reference_seq”字段可以是單個字符（“A”、“T”、“C”或“G”），也可以是多個逗號分隔的字符（例如“TTTT,TTC,GGGGGC”）。 這些字段可以位於 V9 的末尾或中間的任何位置。

V1 V2    V3    V4    V5     V9
9 dbSNP SNV 10007 10007 ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T
9 dbSNP SNV 10009 10009 ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A
9 dbSNP SNV 14824990 14824990 ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP_154:rs140144319;evidence_values=Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD;global_minor_allele_frequency=1|0.004193|211

我最初想到了一個帶有多個分隔符的 awk -F '{print }' 但很快意識到這不是一個可行的解決方案，因為字段之間的行不一致。 dplyr::separate 在這里也沒有真正適應。

我試圖將每個單個字段分別提取到一個新列中，但該命令不處理字段位於行尾的情況：

gsub("Reference_seq[=]([^.]+)[;].*", "\\1", df$V9)

我找不到 grep 的解決方案，只有捕獲組 1 中的字段，如果沒有“;”則停止下列的。 謝謝你的幫助。

Answer 1

由於這有點復雜，我發現使用 stringr 包的str_extract() function 在這里更容易使用。

在這種情況下，我使用 3 行單獨的行來提取感興趣的文本。 並使用 (?<=) 向后看運算符以避免前導文本。

df<- read.table(header=TRUE, text="V1 V2    V3    V4    V5     V9
9 dbSNP SNV 10007 10007 ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T
9 dbSNP SNV 10009 10009 ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A
9 dbSNP SNV 14824990 14824990 ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP_154:rs140144319;evidence_values=Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD;global_minor_allele_frequency=1|0.004193|211"
)

library(stringr)
str_extract(df[,"V9"], "(?<=Variant_seq=).+?;")
str_extract(df[,"V9"], "(?<=Reference_seq=).+?;")
str_extract(df[,"V9"], "(?<=Dbxref=).+?;")

data.frame(Variant_seq, Reference_seq, Db_ref)
#   Variant_seq Reference_seq                  Db_ref
# 1          C;          <NA> dbSNP_154:rs1449034754;
# 2        C,G;          <NA> dbSNP_154:rs1587255763;
# 3 GGGC,CCCCG;            C;  dbSNP_154:rs140144319;

這個最終的數據框現在可以cbind回原來的

Answer 2

你可以做

stringr::str_match(df$V9, "Reference_seq=([^;]+);")[, 2L]
stringr::str_match(df$V9, "Variant_seq=([^;]+);")[, 2L]
stringr::str_match(df$V9, "Dbxref=([^;]+);")[, 2L]

Output

> stringr::str_match(df$V9, "Reference_seq=([^;]+);")[, 2L]
[1] NA  NA  "C"
> stringr::str_match(df$V9, "Variant_seq=([^;]+);")[, 2L]
[1] "C"          "C,G"        "GGGC,CCCCG"
> stringr::str_match(df$V9, "Dbxref=([^;]+);")[, 2L]
[1] "dbSNP_154:rs1449034754" "dbSNP_154:rs1587255763" "dbSNP_154:rs140144319" 

Answer 3

一種可能的解決方案是在與sub配對的字符串上使用strsplit作為子字符串

cbind(df1, sapply(c("Variant","Reference_seq","Dbxref"), function(str) 
  sapply(strsplit(df1[,"V9"],";"), function(x) sub("dbSNP_.*:","",x[grep(str,x)]))))
  V1    V2  V3       V4       V5
1  9 dbSNP SNV    10007    10007
2  9 dbSNP SNV    10009    10009
3  9 dbSNP SNV 14824990 14824990
                                                                                                                                                                                                                                                              V9
1                                                                                                                                                              ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T
2                                                                                                                                                                   ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A
3 ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP_154:rs140144319;evidence_values=Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD;global_minor_allele_frequency=1|0.004193|211
                 Variant   Reference_seq              Dbxref
1          Variant_seq=C Reference_seq=T Dbxref=rs1449034754
2        Variant_seq=C,G Reference_seq=A Dbxref=rs1587255763
3 Variant_seq=GGGC,CCCCG Reference_seq=C  Dbxref=rs140144319

數據

df1 <- structure(list(V1 = c(9L, 9L, 9L), V2 = c("dbSNP", "dbSNP", "dbSNP"
), V3 = c("SNV", "SNV", "SNV"), V4 = c(10007L, 10009L, 14824990L
), V5 = c(10007L, 10009L, 14824990L), V9 = c("ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T", 
"ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A", 
"ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP_154:rs140144319;evidence_values=Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD;global_minor_allele_frequency=1|0.004193|211"
), new = list(c("Variant_seq=C", "Dbxref=rs1449034754", "Reference_seq=T"
), c("Variant_seq=C,G", "Dbxref=rs1587255763", "Reference_seq=A"
), c("Reference_seq=C", "Variant_seq=GGGC,CCCCG", "Dbxref=rs140144319"
))), row.names = c(NA, -3L), class = "data.frame")

Answer 4

這會將 V9 的所有子字段提取到單獨的列中，而無需使用正則表達式或包。 它使用 paste 和 chartr 將 V9 轉換為 dcf 格式，然后使用 read.dcf 將其讀入。 最后我們 append 將創建的列傳遞給 DF。

m <- DF$V9 |>
  paste(collapse = "\n\n") |>
  chartr(old = "=;", new = ":\n") |>
  textConnection() |>
  read.dcf()

DF2 <- cbind(DF, m)

> str(DF2)
'data.frame':   3 obs. of  14 variables:
 $ V1                           : int  9 9 9
 $ V2                           : chr  "dbSNP" "dbSNP" "dbSNP"
 $ V3                           : chr  "SNV" "SNV" "SNV"
 $ V4                           : int  10007 10009 14824990
 $ V5                           : int  10007 10009 14824990
 $ V9                           : chr  "ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T" "ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A" "ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP"| __truncated__
 $ ID                           : chr  "1" "2" "30545117"
 $ Variant_seq                  : chr  "C" "C,G" "GGGC,CCCCG"
 $ Dbxref                       : chr  "dbSNP_154:rs1449034754" "dbSNP_154:rs1587255763" "dbSNP_154:rs140144319"
 $ evidence_values              : chr  "Frequency,TOPMed" "Frequency" "Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD"
 $ Reference_seq                : chr  "T" "A" "C"
 $ clinical_significance        : chr  NA NA "benign"
 $ ancestral_allele             : chr  NA NA "C"
 $ global_minor_allele_frequency: chr  NA NA "1|0.004193|211"

或者這樣寫：

cbind(
  DF, 
  read.dcf(textConnection(chartr("=;", ":\n", paste(DF$V9, collapse = "\n\n"))))
)

筆記

可重現形式的輸入 DF。

DF <- structure(list(V1 = c(9L, 9L, 9L), V2 = c("dbSNP", "dbSNP", "dbSNP"
), V3 = c("SNV", "SNV", "SNV"), V4 = c(10007L, 10009L, 14824990L
), V5 = c(10007L, 10009L, 14824990L), V9 = c("ID=1;Variant_seq=C;Dbxref=dbSNP_154:rs1449034754;evidence_values=Frequency,TOPMed;Reference_seq=T", 
"ID=2;Variant_seq=C,G;Dbxref=dbSNP_154:rs1587255763;evidence_values=Frequency;Reference_seq=A", 
"ID=30545117;Reference_seq=C;clinical_significance=benign;Variant_seq=GGGC,CCCCG;ancestral_allele=C;Dbxref=dbSNP_154:rs140144319;evidence_values=Frequency,1000Genomes,ESP,Phenotype_or_Disease,ExAC,TOPMed,gnomAD;global_minor_allele_frequency=1|0.004193|211"
)), class = "data.frame", row.names = c(NA, -3L))