Trimmomatic error

[Saeideh]

I have a file containing reads. It seems the adapters were removed previously. So I do not need to eliminate the adapters.

When I use fastqc, I have overrepresented sequences in first 7 positions and I want to trim them. That's it.

I wrote this command:

java -jar /usr/local/bin/trimmomatic-0.33.jar SE -phred64 inputfile.fq.gz output.fq.gz HEADCROP:7

Is it wrong?

I get this error:

Exception in thread "main" java.io.IOException: Not in GZIP format
at java.util.zip.GZIPInputStream.readHeader(GZIPInputStream.java:154)
at java.util.zip.GZIPInputStream.<init>(GZIPInputStream.java:75)
at java.util.zip.GZIPInputStream.<init>(GZIPInputStream.java:85)
at org.usadellab.trimmomatic.util.ConcatGZIPInputStream$GZIPHelperInputStream.<init>(ConcatGZIPInputStream.java:109)
at org.usadellab.trimmomatic.util.ConcatGZIPInputStream$GZIPHelperInputStream.<init>(ConcatGZIPInputStream.java:105)
at org.usadellab.trimmomatic.util.ConcatGZIPInputStream.nextGzipInputStream(ConcatGZIPInputStream.java:37)
at org.usadellab.trimmomatic.util.ConcatGZIPInputStream.<init>(ConcatGZIPInputStream.java:16)
at org.usadellab.trimmomatic.fastq.FastqParser.parse(FastqParser.java:132)
at org.usadellab.trimmomatic.TrimmomaticSE.process(TrimmomaticSE.java:192)
at org.usadellab.trimmomatic.TrimmomaticSE.run(TrimmomaticSE.java:285)
at org.usadellab.trimmomatic.Trimmomatic.main(Trimmomatic.java:40)

How can I do it correctly?

Thanks in advance

[GenoMax]

Error is indicating that your input sequence file is not in GZIP format. Are you also sure the input sequence is in phred+64 format?

[dpryan]

In addition to what Genomax wrote, there are some library types (e.g., RNAseq) that commonly show bias at the beginning of the reads. This shouldn't necessarily be trimmed off.

[Saeideh]

I want to scream of happiness.. Thank you guyssss.

Last time, I just renamed the file from .fq to .fq.gz :| :| :| . When GenoMax told me the file is not in GZIP format I did: gzip inputfile.fq and then I got the result.

I'm not sure my input file is in phred+64 format or not. How should I understand it?

dpryan, I know it's ok to have some bias at the beginning of reads and it is because of adapters, but I looked for typical Illumina adapters in my reads files using grep and non of reads have those adapters. So I understood, the adapters might be trimmed before (someone else passed the files to me) and I thought abnormality should be due to contamination and I decided to remove the first 7 bases.

Thank you agaaaaain

[dpryan]

The bias isn't due to primers, it's due to random hexamer priming being biased.

[GenoMax]

I'm not sure my input file is in phred+64 format or not. How should I understand it?

If your data is recent vintage it is unlikely that it is in phred+64 (solexa) format. That said you can test the quality score format of your data by using testformat.sh utility from BBMap suite like this:

$ testformat.sh in=seq.fq

[usad]

you can also just run trimmomatic without the phred switch and it will try to autodetect quality scores (since 0.32)

Best Wishes
björn