SRA submissions

Table of Contents

authors

Noah Hoffman (ngh2@uw.edu)
Chris Rosenthal (crosenth@uw.edu)

package dependencies

biopython
bioscons
csvkit

Introduction

Official walkthrough: http://www.ncbi.nlm.nih.gov/books/NBK47529/

Main submission page: https://submit.ncbi.nlm.nih.gov/

There are three parts to submitting and experiment: the Bioproject, the Biosamples and the SRA fastq sequence submissions.

To start, you will need to use a 3rd party login to sign into the std_crc NCBI account:

https://ncbiinsights.ncbi.nlm.nih.gov/my-ncbi-login-transition-tips/

After you have set up a 3rd party login it will need to be linked to the std_crc NCBI account. Email the std_crc account administrator to add your login to the account.

NOTE:

At any point during the submission process if there are any issues the SRA staff is good at answering questions. Just send an email to sra@ncbi.nlm.nih.gov to have someone assigned to help through the process.

Data

A submission requires access to the datafiles (can be zipped) along with at least two identifying columns of annotation. For this example the two unique, identifying colums are specimen and manuscript_id:

specimen | manuscript_id
------------------------
p7z1tr10 | S1
p7z2tr10 | S2
p7z2tr12 | S3
...

Using the specimen name you can deduce the filenames will look something like if they were pyrosequenced:

p7z1tr10.fastq.bz2
p7z2tr10.fastq.bz2
p7z2tr12.fastq.bz2

Or have two paired-end sequence files from the Illumina Miseq there will be two fastq files. In this example the unique, identifying columns are specimen and label:

specimen     | label
------------------------
m47n701-s502 | 901-2
m47n701-s505 | 901-4
m47n701-s507 | 901-6
...

m47n701-s502_S1_L001_R1_001.fastq.gz, m47n701-s502_S1_L001_R2_001.fastq.gz
m47n701-s505_S1_L001_R1_001.fastq.gz, m47n701-s505_S1_L001_R2_001.fastq.gz
m47n701-s507_S1_L001_R1_001.fastq.gz, m47n701-s507_S1_L001_R2_001.fastq.gz
...

All fastq data files are located in the run folders found here:

/fh/fast/fredricks_d/bvdiversity/data

We have developed scripts to grab these fastq files so you do not need to go hunting.

Now we are ready to begin the three submission walkthroughs.

Bioproject

https://submit.ncbi.nlm.nih.gov/subs/bioproject/

You can start with the either the Bioproject or Biosample submission walkthroughs but the best place to start is with the Bioproject. The Bioproject is considered the "umbrella" point for the Biosamples. Click on 'New submission' and follow the walkthrough. The process is fairly straight forward. For example annotations from previous sumbmissions here are two examples one pyrosequencing and one miseq:

https://www.ncbi.nlm.nih.gov/bioproject/PRJNA420364
https://www.ncbi.nlm.nih.gov/bioproject/PRJNA529191

There will be a step that asks for the Biosample accessions. You will not have the Biosample accessions at this point and there will be too many to submit at this point anyway. Just continue on with the Bioproject submission process skipping this step. After finishing the walkthrough a Bioproject accession will be assigned. Make note of the Bioproject accession and continue one to the Biosample submission process.

NOTE: If you have any questions or need to changes just email sra@ncbi.nlm.nih.gov or bioproject@ncbi.nlm.nih.gov.

Biosample

https://submit.ncbi.nlm.nih.gov/subs/biosample/

Create a 'New submission' Biosample and follow the walkthrough. You will be asked to submit a tab-delimited file with the sample names filled out and other required informatiom. There is a pre-filled template file located in the template folder that can be used:

/fh/fast/fredricks_d/bvdiversity/sra/template/MIMS.me.human-vaginal.5.0.tsv

Using this file the following script will automatically put everything together:

bin/biosample.py --outdir output/CID_NGU_NCBI_SEQ_Submission_FINAL --max-rows 1000 data/CID_NGU_NCBI_SEQ_Submission_FINAL.tsv sample_name,study_id templates/MIMARKS.survey.human-associated.5.0.tsv PRJNA637612

sample_name must be in form mXXnXXXsXXX

For an explanation script options:

bin/biosample.py --help

NOTE: NCBI is enforcing a 1,000 row Biosample batch limit. If you are attempting more than 1,000 samples they will be split into multiple files for multimple Biosample submissions.

Output will look something like this:

sample_name     | sample_title | bioproject_accession | organism                 | host         | collection_date | geo_loc_name       | lat_lon       | ref_biomaterial | rel_to_oxygen | samp_collect_device | samp_mat_process | samp_size | source_material_id | description | label
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
m47n701-s502    |              | PRJNA529191          | human vaginal metagenome | Homo sapiens | missing         | "USA: Seattle, WA" | not collected |                 |               |                     |                  |           |                    |             | 901-2
m47n701-s505    |              | PRJNA529191          | human vaginal metagenome | Homo sapiens | missing         | "USA: Seattle, WA" | not collected |                 |               |                     |                  |           |                    |             | 901-4
m47n701-s507    |              | PRJNA529191          | human vaginal metagenome | Homo sapiens | missing         | "USA: Seattle, WA" | not collected |                 |               |                     |                  |           |                    |             | 901-6
...

Upload the table as part of the biosample submission process. After some time accession numbers for each sample_name will be generated. When they are ready there will be an attributes file for you to download via the main biosample submission portal page. We will use this attributes file to fill out the final sra sequence upload template.

NOTE: From time to time NCBI will update the MIMARKS.survey.human-vaginal template which will force us to update our own. To get the latest template go to:

https://submit.ncbi.nlm.nih.gov/biosample/template/

Under "GSC MIxS packages for genomes, metagenomes, and marker sequences" select "MIMARKS Survey related" -> "human-vaginal" -> "Download TSV". Place the tsv file into the templates folder adding any additional required annotation and update this README accordingly.

Sequence Upload

https://submit.ncbi.nlm.nih.gov/subs/sra/

This is the last step in the process. Click on the 'New submission' tab above and follow the steps. There will be an sra submission form to fill out and submit. Here is where we upload the fastq data files. Using your Bioproject accession (or if you submitted the Bioproject first it will already be included in your biosample_accession attributes file) run the bin/sra_meta.p script. Note, the same upload limit of 1,000 samples applies here so you may need to go through multiple SRA submissions wizards to finish the submission:

bin/sra_meta.py --outdir output/Overbaugh_NCBISRA/fastq --out output/Overbaugh_NCBISRA/SRA_metadata_acc.tsv data/Overbaugh_NCBISRA.tsv output/Overbaugh_NCBISRA/attributes.tsv templates/SRA_metadata_acc.tsv ../data

Or if a miseq sample then use this script:

bin/sra_meta_miseq.py --outdir output/LancetHIV_Kelleretal_2019/fastq --out output/LancetHIV_Kelleretal_2019/SRA_metadata_acc.tsv output/LancetHIV_Kelleretal_2019/attributes.tsv templates/SRA_metadata_acc_miseq.tsv ../data

The fastq files will be gathered and placed in the --outdir folder while the filled out SRA form will placed in --out. The form file will have one row per specimen/sample which will look somewhat like this:

biosample_accession  | title | library_ID | design_description                                                                                                                                                                                                                                                                                    | library_strategy | library_source | library_selection | library_layout | platform | instrument_model    | filetype | filename1 
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
SAMN04859440         | S1    | p7z1tr10   | DNA was extracted using the Bacteremia Kit (Mobio). The V3-V4 region of the 16S rRNA gene was targeted for broad-range PCR with pyrosequencing. 6-bp barcodes were used with the reverse primer to facilitate multiplexing. Reactions were purified using Agencourt AMPure beads prior to sequencing. | AMPLICON         | METAGENOMIC    | PCR               | single         | _LS454   | 454 GS FLX Titanium | fastq    | p7z1tr10.fastq.bz2
...

Go ahead and upload the filled out form SRA --out file when asked for it.

The last step in this process will have the ftp upload instructions. They will provide an ftp username, password and upload location. You will also need to create a folder to upload the files to. Use the following script with the parameters provided from NCBI as follows:

bin/ftp_put.py ftp-private.ncbi.nlm.nih.gov subftp w4pYB9VQ uploads/ngh2@uw.edu_u25A5oa4 LancetHIV_Kelleretal_2019 output/LancetHIV_Kelleretal_2019/fastq

The username, password and upload folder will change periodically so make sure to use the latest, correct parameter information.

Finally

When you get the specimen accessions create another column called 'sequence_accession' in the data sheet (data/Gorgos_Sycuro_SDC_Table_S1.tsv) and manually enter the sequence accessions. Write an email to Sujatha giving her the updated data sheet with the specimen accessions and also she will need the project accession

Congratulations, you have completed the sra submission process!

crosenth / sra_submission