Submitting sequences to SRA

[morning latte]

Hello experts,

I have spent whole day trying to upload my Illumina paired-end reads to the NCBI SRA. I have read through the manual over and over but it is quite confusing. Could anyone help me figure out how to solve my problems? Any advice would be greatly appreciated.

I have created one BioProject and one BioSample. The BioSample contains 8 samples.

1. When I create a new submission in the SRA, should I create an experiment and a run for each sample? Or should I create only one experiment and create 8 different runs under one experiment?

2. For Illumina paired-end reads, can I upload an interleaved file for each sample or should it be two separate files (R1 and R2)?

3. I tried to upload two separate files (R1 and R2) which were quality trimmed for one sample. The upload failed. Should I upload only raw sequences or assembled contigs?

[yzzhang]

for the question 2, you can upload the interleaved file. It is acceptable.

[mikep]

The SRA is arcane.

biosample refers to the source of your material. If it was 8 different people then that is 8 different biosamples. if it was, say, 2 people which were given a time course of treatment (for 8 experiments) that's 2 biosamples

Each will then be a different experiment. you only create multiple runs when you quite literally have run exactly the same sample/experiment multiple times, or for instance you are uploading lanes separately.

Most of the time when I use sra, it's 1 biosample, 1 experiment, 1 run per file (or pairs of files for paired end).

You can use their spreadsheet to upload multiple experiments at a time (but you need to have created the biosamples first)

In what way did the upload fail?

[morning latte]

Thanks mike for your helpful advice! I very agree with you that the SRA is arcane.