Paired-end mode : Read1 quality revision

[hubert93]

I've trimmed nexteraXT adapter for paired-end illumina data, using this command.

skewer -x CTGTCTCTTATACACATCT sample_1.fastq.gz sample_2.fastq.gz

And i realized that the quality of trimmed read1 is revised to better score.
For example,

Raw read1)
@Miseq:70:000000000-AWLU0:1:1101:14472:1001 1:N:0:TAGGCATGAAGGAGTA
NTGG ... ACCAGACGGTGAGTAACGAAAGCCAGGGGCGCGAACCGGATTAGAAACCCTTGTAGTCCCTGTCTCT
+
#8AC ... 3:C=>5ACGGG>GGFF>7=DGB47CCF7CC@>BD))9ED<C=<6@7CF6C#################

Trimmed read1)
@Miseq:70:000000000-AWLU0:1:1101:14472:1001 1:N:0:TAGGCATGAAGGAGTA
NTGG ... ACCAGACGGTGAGTAACGAAAGCCAGGGGCGCGAACCGGATTAGAAACCCTTGTAGTCC
+
#8AC ... GGGGGFGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGFFGGGGGGGCCCCC

I assume that quality of read1 is corrected by being trimmed adapter sequence using paired-end information. I wonder it is true, and whether there is any way to prevent it.