Trouble with box-plot of quality data in SOLID

[rivabros]

hi everyone,
I have used the SOLiD2std.pl to change the csfasta and qual files to standard fastq files.
I then ran fastqc to view boxplots of quality data and got these results, attached.
They seem to have poor quality every 5 nt, as if the primer 5 of the procedure failed....
Has anyone seen this type of quality plots before???
I am new with solid data, have worked with ilumina and 454 before.
Thanks for any guidance in advance.
cheers
maximo

[Jean]

You can not convert colorspace to basespace directly. See here why:

http://seqanswers.com/forums/showpost.php?p=59156&postcount=4

[colindaven]

An alternative route to using FastQC, which is very nice, is to align first and use the resulting BAM file as input.

Another program I have looked at for duplicate analysis is prinseq, which also offers graphical output.

[Krish_143]

Did you removed adapter info or trimmed your reads.
what about other fastQC results they are pass or fail.

[rivabros]

Did you removed adapter info or trimmed your reads.
what about other fastQC results they are pass or fail.

Some aspects passed and others failed,
I did NO removal of adapters.... Not sure if it was done when i received the csfasta and qual files..
Taking into consideration that this is a direct conversion from color to base space, its understandable that there may be confusing results.... (In all about 45% of reads mapped with color-space-novoAlign to a reference, and only 15% usnig colorspace-tophat/bowtie...)
Any thoughts?
PASS Basic Statistics Corrida_4_FC_1_01_01CVATFR001_F3.fastq
FAIL Per base sequence quality Corrida_4_FC_1_01_01CVATFR001_F3.fastq
WARN Per sequence quality scores Corrida_4_FC_1_01_01CVATFR001_F3.fastq
WARN Per base sequence content Corrida_4_FC_1_01_01CVATFR001_F3.fastq
FAIL Per base GC content Corrida_4_FC_1_01_01CVATFR001_F3.fastq
FAIL Per sequence GC content Corrida_4_FC_1_01_01CVATFR001_F3.fastq
WARN Per base N content Corrida_4_FC_1_01_01CVATFR001_F3.fastq
PASS Sequence Length Distribution Corrida_4_FC_1_01_01CVATFR001_F3.fastq
PASS Sequence Duplication Levels Corrida_4_FC_1_01_01CVATFR001_F3.fastq
PASS Overrepresented sequences Corrida_4_FC_1_01_01CVATFR001_F3.fastq
FAIL Kmer Content Corrida_4_FC_1_01_01CVATFR001_F3.fastq

[colindaven]

I would recommend using SAET to clear up any colour errors before alignment. Normally I get 1-5% more aligned reads using this (apparently a lot more with ECC data).

I don't think your fastqc boxplot results are too bad, but I've seen better ones.