1、直接測試

root@PC1:/home/test# ls
a.fna
root@PC1:/home/test# cat a.fna
>NC_019458.2 Ovis aries breed Texel chromosome 1, Oar_v4.0, [whole genome shotgun sequence] ## 利用sed實現僅保留紅色背景部分
gataaaaaataaatagaaacaaaatcactgaagaaCCAGTGTGCCTGCTCAGGTCAGATGAAGCCAGAGGGCTGCCAGAG
GGCAAGCGAGCTGCGTTGCCTGGAAAAAGTTAAACACACAGAGAGCATGGTGGCTCTGATACTTTCTAGAAGGATTAAAG
TCACTTTCCCAGTCTTTATGAGAATTGGGCCGAAGCTTAGCTGGTGCAACGAATTTAGAAATGAATGCACTTGCATTTGA
 
>NC_019458.2 Ovis aries breed Texel chromosome 1, Oar_v4.0, [whole genome shotgun sequence]
AGATGATGTGTCTTTGCCTTGAgctaaaaattttagaataatctgaACGTCATCTGAGGAACCTGCTTCTGGCGTGGTTT
TGGTGTCAGCATCTTCTCACCCTCTCTAGTAATTTTCAGTATGCATTTCTATTTTCGTGTAGTTATTTACAGGAGCATTT
TATGGAAAACCGGCTCAAATCTTTTTGGGTGCAGGGGTAGTTCAAATGCACTGAGACCCTCAGTTTCACTTGCTAATCTC
 
>NC_019458.2 Ovis aries breed Texel chromosome 1, Oar_v4.0, [whole genome shotgun sequence]
CTCCAGAAACCCTGTTCTCCTCGAGTGACAAGGTCAGCAGGGCAGCACGTGTGTTCCTGTCACTGCCAACTCAAGAATAT
GAAGTTTAAAGAGTTTCACCATCAAATGCAGTGTCGTGGACTGCCCCTGAACAGGTGTTTATAATCACGTGTGCAAGTGA
AGCAAGCACAAATCCTCAGTGGAAAACGGGCAGAGGACACGAGCagacaattctttttaaaaactgcacaaATTAGCACA
 
>NC_019458.2 Ovis aries breed Texel chromosome 1, Oar_v4.0, [whole genome shotgun sequence]
CTAGGCACGGATGAGCGTGCCTACCGTGTTGCATGGAGGTAACAGATGCCAGAGCCCGGAGGAGGCGCAAAGCTCACAAA
CAGATGCGGACCGCAGGAAGCCGGGACGGCCTTCCTCCCCTGAAGCAGGAGGACGCGCCCTACAGAAAGCCGCTCGATCC
TCCAGGCATTTGTTGTGAGCACTTAATCATCATTCGATCATTTGACGTGTACTCACTAGTAAAAGGCAGGACTGTGTCCC
root@PC1:/home/test# sed 's/\(>[^\s]\+\)\s.*\(\[.*\]\)/\1 \2/' a.fna  ##使用正則表示式，第一個小括號被儲存在\1中，第二個小括號被儲存在\2中
>NC_019458.2 [whole genome shotgun sequence]
gataaaaaataaatagaaacaaaatcactgaagaaCCAGTGTGCCTGCTCAGGTCAGATGAAGCCAGAGGGCTGCCAGAG
GGCAAGCGAGCTGCGTTGCCTGGAAAAAGTTAAACACACAGAGAGCATGGTGGCTCTGATACTTTCTAGAAGGATTAAAG
TCACTTTCCCAGTCTTTATGAGAATTGGGCCGAAGCTTAGCTGGTGCAACGAATTTAGAAATGAATGCACTTGCATTTGA
>NC_019458.2 [whole genome shotgun sequence]
AGATGATGTGTCTTTGCCTTGAgctaaaaattttagaataatctgaACGTCATCTGAGGAACCTGCTTCTGGCGTGGTTT
TGGTGTCAGCATCTTCTCACCCTCTCTAGTAATTTTCAGTATGCATTTCTATTTTCGTGTAGTTATTTACAGGAGCATTT
TATGGAAAACCGGCTCAAATCTTTTTGGGTGCAGGGGTAGTTCAAATGCACTGAGACCCTCAGTTTCACTTGCTAATCTC
>NC_019458.2 [whole genome shotgun sequence]
CTCCAGAAACCCTGTTCTCCTCGAGTGACAAGGTCAGCAGGGCAGCACGTGTGTTCCTGTCACTGCCAACTCAAGAATAT
GAAGTTTAAAGAGTTTCACCATCAAATGCAGTGTCGTGGACTGCCCCTGAACAGGTGTTTATAATCACGTGTGCAAGTGA
AGCAAGCACAAATCCTCAGTGGAAAACGGGCAGAGGACACGAGCagacaattctttttaaaaactgcacaaATTAGCACA
>NC_019458.2 [whole genome shotgun sequence]
CTAGGCACGGATGAGCGTGCCTACCGTGTTGCATGGAGGTAACAGATGCCAGAGCCCGGAGGAGGCGCAAAGCTCACAAA
CAGATGCGGACCGCAGGAAGCCGGGACGGCCTTCCTCCCCTGAAGCAGGAGGACGCGCCCTACAGAAAGCCGCTCGATCC
TCCAGGCATTTGTTGTGAGCACTTAATCATCATTCGATCATTTGACGTGTACTCACTAGTAAAAGGCAGGACTGTGTCCC

2、簡單示例

root@PC1:/home/test# echo "hello world" | sed 's/\(hello\).*/world \1/'  ##hello儲存在\1中，匹配了完整的hello world，因此替換為 world hello
world hello

3、示例3

root@PC1:/home/test# cat mysed.txt    ## 在第一個Beijing和第4個Beijing後面新增2008
Beijing Beijing Beijing Beijing
London London London London
root@PC1:/home/test# sed 's/\(^Beijing\)\(.*\)\(Beijing$\)/\1 2008\2\3 2008/' mysed.txt 
Beijing 2008 Beijing Beijing Beijing 2008
London London London London

參考：http://c.biancheng.net/linux/sed.html