linux - GREP 或 AWK : Search in the first N characters of each line, 并输出匹配模式的周围行

Question

我有一个 RNA-seq 数据，如下所示:

@J00157:85:HNNJLBBXX:5:1101:2869:15047 1:N:0:ATTACTCG+TATAGCCT
CGACGCTCTTCCGATCTGAGCTGCAGCCTCGGCCCCAGGATCCCCCTGGGGGACTGGACGCTGCTATTGATTCACGAGGCGCTCAGATCGGAAGAGCACAC
+
AAFFFJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJFJJJJJJJJJFJJJJJJFJJJJJJJJFJJJFJFJJJJJJJJJJJJJJJJ
--
@J00157:85:HNNJLBBXX:5:1101:12550:15574 1:N:0:ATTACTCG+TATAGCCT
GCTCTTCCGATCTGCTATTGATGACTGTCCTCTGTTCTTTCTTTCACAGTAGACGAGGACAGATCGGAAGAGCACACGTCTGAACTCCAGTCACATTACTC
+
AAAFFJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
--

如果我们把@后面的所有内容都当成一个section，可以看到只有第二行才是真正的测序信息，1,3,4,5是logistic/quality信息。

目标是提取每行前N(N=35)个字符中包含“GCTGCA”的序列(第二行信息)，并且同时输出周围的行(前面1行，匹配行后面3行)。

一个示例答案是

@J00157:85:HNNJLBBXX:5:1101:2869:15047 1:N:0:ATTACTCG+TATAGCCT
CGACGCTCTTCCGATCTGAGCTGCAGCCTCGGCCCCAGGATCCCCCTGGGGGACTGGACGCTGCTATTGATTCACGAGGCGCTCAGATCGGAAGAGCACAC
+
AAFFFJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJFJJJJJJJJJFJJJJJJFJJJJJJJJFJJJFJFJJJJJJJJJJJJJJJJ
--

我试过的是

awk 'substr($0, 1, 35) ~ "GCTGCA"' filename.fastq > newfile.fastq
grep -B 1 -A 2 -E GCTGCA filename.fastq > newfile.fastq
awk '{a[++i]=$0;}{substr(a[++i], 1, 35) ~ "GCTGCA"}{for(j=NR-1;j<=NR+2;j++)print a[j];}' filename.fastq > newfile.fastq

第一个不能输出周围的线。第二个不能限制每行前 35 个字母的模式匹配。第三行应该可以工作，但它给了我有线输出(这显然是不正确的):

@J00157:85:HNNJLBBXX:5:1101:14235:1367 1:N:0:ATTACTCG+TATAGCCT
@J00157:85:HNNJLBBXX:5:1101:14235:1367 1:N:0:ATTACTCG+TATAGCCT
TCTNCTCTTCCGATCTACCCCACACACCCCCGCCGCCGCCGCCGCCGCCGCCCTCCGACGCACACCACACGCGCGCGCGCGCGCGCCGCCCCCGCCGCTCC
TCTNCTCTTCCGATCTACCCCACACACCCCCGCCGCCGCCGCCGCCGCCGCCCTCCGACGCACACCACACGCGCGCGCGCGCGCGCCGCCCCCGCCGCTCC
+
+
AAF#FJJJJJJJJFJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJFJFJJJJJJJJJJJJJJJJJJJJJFJJJJJJJJJJJJJJJFJJAJJJJJFJJJJ7JJ
AAF#FJJJJJJJJFJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJFJFJJJJJJJJJJJJJJJJJJJJJFJJJJJJJJJJJJJJJFJJAJJJJJFJJJJ7JJ
--
--

Answer 1

使用 gawk 多字符 RS 支持。

awk -v RS='\n--' -F'\n' 'substr($2,0,35)~"GCTGCA"{print $0 RS}' file

您使用记录分隔符定义记录。