SRA Metadata and Submission Overview

Introduction

The SRA metadata describes the technical aspects of sequencing experiments: the sequencing libraries, preparation techniques and data files.

Most of descriptive information is captured at the level of the SRA EXPERIMENT and will be displayed in the public record. It is therefore imperative that submitters provide clear and informative Title and Description for each EXPERIMENT.

EXPERIMENT as viewed on the NCBI SRA public site

Anatomy of the SRA data

Organizational framework of the SRA data is based on the concepts of STUDY (corresponds to PROJECT in the BioProject database), SAMPLE (corresponds to SAMPLE in the BioSample database), EXPERIMENT, and RUN.

Relations between SRA objects, data, and submissions

The SRA publicly accessioned objects are STUDY (accession in the form of SRP#), SAMPLE (SRS#), EXPERIMENT (SRX#), RUN (SRR#). SUBMISSION has a non-public accession in the form of SRA#.

The SRA EXPERIMENT and RUN objects contain instrument and library information and are directly associated with sequence data.

SRA data pertaining to a STUDY can be deposited in more than one SUBMISSION.

A SAMPLE can be shared between STUDIES and between SUBMISSIONS.

SRA metadata: EXPERIMENT

Each SRA EXPERIMENT (SRA accession SRX#) is a unique sequencing result for a specific sample.

Example

Six sequencing libraries were prepared from a single biological sample. Three were single-end libraries, and three paired-end, although the paired-end libraries were sequenced using both paired and unidirectional sequencing. Two of the single-end libraries were treated using a targeted selection approach for some runs. Libraries were sequenced on two different instruments at three sequencing labs. In all there are 13 different combinations of library + sequencing strategy + layout + instrument model. Each combination represents a unique EXPERIMENT.

Additional information may be included in the EXPERIMENT. For example, you should differentiate biological replicates using EXPERIMENTs if sequencing results were obtained separately from each animal in a group of otherwise identical animals (treated, non-treated, healthy, infected, etc.), the above EXPERIMENTs may be represented by a combination replicate number + library + sequencing strategy + layout + instrument model.

An SRA EXPERIMENT is the main publishable unit in the SRA database.

Most of descriptive information is captured at the level of the SRA EXPERIMENT will be displayed in the public record.

Linking metadata and data: RUN

SRA RUN is simply a manifest of data file(s) that are derived from sequencing a library described by the associated EXPERIMENT.

When submitting in The SRA Submission Portal wizard, submitters only provide types and names for the sequence data files that they will be uploading.

Paired-end data files (forward/reverse) must be listed together in the same RUN in order for the two files to be correctly processed as paired-end.

All data files listed in a RUN will be merged into a single *.sra* archive file. Therefore, files from different samples or experiments should not be grouped in the same RUN.

Contact SRA staff

Please email sra@ncbi.nlm.nih.gov.