NCBI_sequence_Read_Archive

[Luyi Tian]

Hi, I am a new comer, and I am working on RNA-seq these days and want to further analyse the data produced in previous study. But I am confused by the data in the NCBI.

The file size varies a lot in the same species. some only about 500mb, some could up to 10Gb. Why are there such a difference? These difference even occur in the different run in same sample. for example in this http://www.ncbi.nlm.nih.gov/sra/SRX000571?report=full. the data in two runs is totally different:

# Run # of Spots # of Bases
1. SRR002321 54,856,271 2G
2. SRR002323 14,761,931 531.4M



I choose the bigger one SRR002321 and fed it into tophat using command :
$ tophat -p 8 ./bowtie/bowtie-0.12.5/indexes/genome SRR002321.fastq
( the ../indexes/genome is from NCBI build 37.2)
but failed this is the error code:
Error: segment-based junction search failed with err =-9
any guys know what ''err= -9'' means?


Then I tried another file SRR002320 in http://www.ncbi.nlm.nih.gov/sra/SRX000605 ( this file and SRR002321 are in the same study) with the same code:
$tophat -p 8 ./bowtie/bowtie-0.12.5/indexes/genome SRR002320.fastq
this time everything is ok.
I checked the log and found there was less than 50% seqences been aligned to the reference genome. Here is the text in bowtie.left_kept_reads.fixmap.log:
# reads processed: 38255195
# reads with at least one reported alignment: 17219294 (45.01%)
# reads that failed to align: 20387248 (53.29%)
# reads with alignments suppressed due to -m: 648653 (1.70%)
Reported 35237526 alignments to 1 output stream(s)
Anyone knows how to improve the results?

[Bukowski]

I'm going to hazard a guess that the sample data has been generated across 2 sequencer runs, generating 2Gb in one run and 500Mb in the other. Hence the total sample size is 2.5GB which is correctly reported on the same page.

As for differences of size of RNA-Seq samples between experiments, this could be due to anything - it's going to change dependent on the read depth you hope to achieve, which platform you use etc. There's no standard 'output size' for an RNA-Seq sample