SRA to fastq

Question

[vimlakany]

The command used to convert sra to fastq is fastq-dump --split-3 ERR738423.sra The above sra is single-end data. SRA file size is 2.2GB; using fastq-dump command fastq file obtained is 10.2GB; in ENA fastq file is 7GB. Why there is a huge difference in size?

Answer 1

[Satyajeet Khare]

Hi,

ENA files are smaller that GEO because in line 3 '+' character is not followed by the sequence identifier. For example, in GEO it is

@ENTIRE_SEQ_ID
GATTTGGGGTTCAAAGCAGTATCGATCAAATAGTAAATCCATTTGTTCAACTCACAGTTT
+ENTIRE_SEQ_ID
!''*((((***+))%%%++)(%%%%).1***-+*''))**55CCF>>>>>>CCCCCCC65

But in ENA it is

@ENTIRE_SEQ_ID
GATTTGGGGTTCAAAGCAGTATCGATCAAATAGTAAATCCATTTGTTCAACTCACAGTTT
+
!''*((((***+))%%%++)(%%%%).1***-+*''))**55CCF>>>>>>CCCCCCC65

At least, that was the case with my datasets.

[vimlakany]

But this problem exits in SRA to fastq also.

[Satyajeet Khare]

Yes. I mean it depends on the source of SRA file. Can you 'head' both fastq files and check if there is difference in line 3?

Answer 2

[shenwei356]

Check the content:

1. Sequence number:

   grep -c '^' reads.fastq
   

2. Sequence header line format:

   grep '^' reads.fastq | head -n 10
   

[vimlakany]

Sequence number: 168124864

Sequence header line @ERR738423.1 HWI-ST365_0182:2:1101:1134:2086#CGATGT length=50 AGTGTCTAAGGGCGCATGGTGGATGCCTTGGCATCGAGAGCCGATGAAGG +ERR738423.1 HWI-ST365_0182:2:1101:1134:2086#CGATGT length=50 @@=D?DDD?DDF1C1FHIGE@GGHGEHHGEIC>B>FHH?AGC>AFHCHGG @ERR738423.2 HWI-ST365_0182:2:1101:1152:2089#CGATGT length=50 CCGAACCCGGAAGCTAAGCCTGCCAGCGCCGATGATACTGCCCCTCCGGG +ERR738423.2 HWI-ST365_0182:2:1101:1152:2089#CGATGT length=50 CCCFFFFFHHGHHJJIIIIJJJJJIJIJIJJGHIJJJJIIIJJFHHFFDD @ERR738423.3 HWI-ST365_0182:2:1101:1095:2121#CGATGT length=50 TCAAGCACACCGCCGAAGCCGCGGCACATCCACCTTGTGGTGGGAGTGGG

Why should we check sequence number and sequence header line format?

[shenwei356]

i mean compare these infomation between the two files.