[Help!] How can I extract high quality reads from the output file of NovoAlign

[qc.share]

Hello Everyone,
Now, we are trying to extract uniquely aligned reads(or high quality reads) from the output file of Novoalign.
There are 5 different reads in this file(please see below), which kind of reads could be used to export the methylation signals(MeDIP-seq)?
In our Service Report, there are 2 kind of reads:
1) Aligned reads:The reads aligned to reference genome with no more than 2 mismatches.
2) Uniquely aligned reads: The reads aligned to the unique location in the genome.
I found that "Uniquely Aligned Reads" have more than 2 mismatches (for example:34G>C 35T>C 36G>T).

@HWUSI-EAS1522_100625:1:120:18868:18135#0/1 S AAGAGGTTTCCAACGAAGGCCTCAAACAGGTCAATG #################################### QC
@HWUSI-EAS1522_100625:1:2:3391:8944#0/1 S AAAGTAAAGTGTGATGAGGCATCGTGAACTGAAACC "0''')',((,33338AAAAAAAAAAAAAAAAAAAA8" U 37 54 >chr22 27508807 F . . . 1G>A 2C>A
@HWUSI-EAS1522_100625:1:120:18870:10840#0/1 S CCATTCCATTCCATTCCACTCCACACCACTCCACTG CCCAACBCCCCCCCCCCCC<CCCCACCCCCCCCCCA R 6
@HWUSI-EAS1522_100625:1:2:3509:1310#0/1 S AGTGAGGACCTGGATCGGCACAGCGCACTGCAGCGC )&)))00000AAAAAAAAAAAAAAAAAAAAAAAAAA U 6 96 >chr22 20772761 R . . .
@HWUSI-EAS1522_100625:1:2:3662:5720#0/1 S AGGACAAGCTGAGATGAATAGTGTTTTGCAAAGGAG "-'&'(3,,//AAAAAAAAAAAAAAAAAAAAAAAAAA" U 40 62 >chr22 33074711 R . . . 34G>C 35T>C 36G>T

So, If I want to choose high quality read to export methylation signal, which kind of read should I use?

Thanks very much, everyone!

qc.share