fastq quality check with Q=33 or Q=64?

[feng]

Hi all,

I'm trying do quality checking with FASTX toolkit,

http://hannonlab.cshl.edu/fastx_toolkit/commandline.html#fastq_quality_filter_usage

Now I'm confused with different fastq. I list some samples as following, Is there anyone know which Q (33/64) I should used for data analysis?

@SRR036482.94 SOLEXA8_1:2:1:0:1735 length=51
AAAATAAAATATTGGCACAAAGAGATTTCAAAAATATCACTGCATCCTATA
+SRR036482.94 SOLEXA8_1:2:1:0:1735 length=51
=HCHHHEC@EHHHHHHHHDCGGBGHHHHHFHHHHHEEHHHHHFHHHHHHHD
@SRR036482.95 SOLEXA8_1:2:1:1:1455 length=51
ATTATAGAAAGCCAGAGGGAACTAATCACTTCAGTGCAACTGCAGCAGGTA
+SRR036482.95 SOLEXA8_1:2:1:1:1455 length=51
HHDDC=C7?D??DDD8ECC0ACGEGHHEDHHEHE@A8GC?:C4;D=CA@53

Many thanks.

[fkrueger]

This definitely looks like Sanger quality values (Phred+33).

See FastQ for more information on the FastQ data format.

[kmcarr]

fkrueger is correct, they are Sanger (Phred+33) encodings. The accession numbers (SRR) indicate that this is data downloaded from the the NCBI SRA. SRA data will always be Phred+33.

[aloliveira]

Feng,


How are you? There is an excelent paper that describes the FASTQ encoding. But kmcarr is correct, all the FASTQ sequences from ncbi are Phred+33.

http://www.ncbi.nlm.nih.gov/pubmed/20015970