Bowtie2 Error: Saw ASCII character 10 but expected 33-based Phred qual

Question

[jaqx008]

Hello Everyone, I encountered an error while mapping with bowtie2. I thought at first something happened to the file during download and I have redownloaded the file and tried the mapping again and each time I get the error below when I run the following command. Can someone help please? Btw I have looked other solutions but the error is either not exactly the same or I dont understand it.

bowtie2 -p 24 --local -x file.index -U sra_data.fastq -S file.sam

Saw ASCII character 10 but expected 33-based Phred qual. libc++abi.dylib: terminating with uncaught exception of type int (ERR): bowtie2-align died with signal 6 (ABRT)

[GenoMax]

Is the data not in sanger fastq format?

You can quickly check using testformat.sh from BBMap suite.

$ testformat.sh in=seq.fq.gz
sanger    fastq    gz    interleaved    150bp

[Bastien Hervé]

Can you give the result of

head sra_data.fastq

Thanks

[jaqx008]

@DRR032678.1.1 1 length=101 
NCGTAAGACATGCAATGAATTGTGTGGTACAAGCTGATGTATGGTCCAGGGTTTGTAGTGCAGCTATCACANAGCGATAGATNATATCAGAATATCGTCCC
+DRR032678.1.1 1 length=101
#4=4=DBBCB?DFFBAGB?CEEFCAC?F@HFEF?FI?FFD9CDD:DDBEFEDFDFF<CCDG@C@@FFDED@#-5A?BBD>BD#,5=?BEDBA@BA@B#### 
@DRR032678.2.1 2 length=101
AGTCCTGTCTGTGGCCAGGGCGTAGCCAGACGCTGAAGCGGGGAAACGTTACGTACTATCTGGACGGCGCCCACACTCAGCGGAGAATCCAGGTGGGCGTG
+DRR032678.2.1 2 length=101 
@@;?DDDDHFFHHIIGIGGIHGFGGGIIGEHGGG0DHEGEHIFB>=ABA39??((55@>CCA<@<7>BBB;>&8?@?>9::95<9&24::@########## 
@DRR032678.3.1 3 length=101 
CTGTGATCGTATGTCAAATGTATGCCCTTTTATGAAAGCTGCCACTTTCATCTGTTAAACTCAATTGTACTTCCAGGTTTGTAACACCCTAACCTGTAGTA

[GenoMax]

$ testformat.sh DRR032678.fastq 
sanger  fastq   raw     single-ended    101bp

Should be in normal sanger fastq format.

[jaqx008]

I tested it and it gave sanger fastq raw single-ended 101bp the other files I obtained were made with sanger too and they mapped fine except this one.

[jaqx008]

Is something wrong with my fastq data?

[Joe]

There doesn’t appear to be anything obviously wrong with the overall structure of the data at a glance, but it could be a single character in the QUAL line of a single read which is causing Bowtie to choke - have you got any log files or information about what particular read might be the issue?

ASCII character 10 is the linefeed character which AFAIK certainly isnt a valid, not least because its not a single character. You may need to search your reads for a ^J character. If these sequences are from the internet, someone may have tried to manually edit them (God knows why) in the past.

The only permissible characters are the ones listed here:

https://support.illumina.com/help/BaseSpace_OLH_009008/Content/Source/Informatics/BS/QualityScoreEncoding_swBS.htm

Edit: ascii 10 is linefeed not backspace, my bad. Same approach applies though.

[jaqx008]

Thanks for your input. I was not able to find anything like H when i seached. However, I did not really need the header information down in my analysis so I was able to go around this by converting the fastq to fasta then mapping with -f in bowtie2. Thanks for your contribution.

[Bastien Hervé]

I would not recommend you to align your read without quality (fasta). Try to split you fastq and re-run the mapping till you find the line involved in this error. Maybe not the whole file is corrupt.