Submitting data to public databases is super important for publically funded laboratories, but it is not always a quick or intuitive process. ncbi-submit provides a simple and repeatable way to upload programmatic submissions to NCBI's SRA and GenBank with shared or unique BioProjects and BioSamples. Data can be uploaded as XML or zip files to either the Test or Production environments, and once there, the reports produced by NCBI can be analyzed to check on submission status and get BioSample accessions.
- ncbi-submit
- Installation:
- Testing
- Usage
- Updating samples that have already been submitted
- Input File Paths Explained
- Links to xml template examples/schema:
To install from PyPI in a virtual environment .venv:
python3 -m venv .venv
. .venv/bin/activate
pip install ncbi-submitTo install from conda (not yet set up) in a new environment ncbi:
conda create -n ncbi ncbi-submitAdd NCBI credentials to file ./.login_credentials or edit them in either:
./example/test.shor./config/config.py
To test creating all example files, run:
./example/test.shThis script ^^^ could also be a good starting point for your own NCBI submission pipelines. Note: There are several blocks of code in there can be commented in/out, as needed.
ncbi_submit is intended for use on the command line, but the class ncbi.NCBI can be imported and used within custom python scripts.
There are three main actions the script can do:
file_prep:- Prepares .tsv & .xml files for SRA, BioSample, & BioProject submissions
- Used to prepare all files for initial submission to NCBI
- To add in biosample accessions and prepare for GenBank submission, include the flag
prep_genbank:- Prepares .zip, .sbt, & .tsv files for GenBank Submission
- Used to add BioSample accessions from a BioSample submission for a GenBank submission
ftpsubmission or checkup:- Interacts with NCBI's ftp host to do either of the following:
submitdata to NCBI databasescheckon previous ftp submissionsget-accessionsfrom all previous ftp submissions
- Interacts with NCBI's ftp host to do either of the following:
example:- Writes out example files for one or both of:
- config.py file (tells
ncbi_submitlots of important info) - template.sbt (used for genbank submission)
- config.py file (tells
- Writes out example files for one or both of:
The required parameters vary by which of the above actions you're attempting but at minimum require a plate and outdir. To limit the number of parameters required via command line, a config file must be used. When running from the command line, one of the three actions (file_prep or ftp) must be specified. With python, these are associated methods you may use on a single NCBI object.
Run this command to get a example config.py file in a directory called './ncbi':
ncbi_submit example --config --outdir "./nbci"A BioProject accession can be created in NCBI's submission portal, but it can also be created by ncbi-submit either as part of a BioSample/SRA submission or all by itself.
Steps for creating a new BioProject accession via ncbi-submit:
- In your
config.pyfile, setbioproject['create_new'] = True - Follow the below file preparation advice
- Follow the below file submission advice, but if you're only creating a BioProject and don't want to submit any other data, you can omit the --fastq_dir and --plate options and specify a --subdir instead (as the name of the directory to be used in NCBI's ftp site)
- Once you have results, add the new accession to your
config.pyfile atbioproject['bioproject_accession']and setbioproject['create_new'] = False
Note: This is the minimum required info for preparing data. Other parameters may be necessary for more functionality or other tasks.
from ncbi_submit import ncbi_submit
ncbi = ncbi_submit.NCBI(
fastq_dir = myFastqDir,
seq_report = mySeqReport,
plate = myPlate,
outdir = myOutdir,
config_file = myConfig,
)
ncbi.write_presubmission_metadata()Shell:
ncbi_submit file_prep \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--seq_report "${SEQ_REPORT}" \
--primer_map "${PRIMER_MAP}" \
--primer_scheme "${SCHEME_VERSION}" \
--outdir "${NCBI_DIR}" \
--gisaid_log "${GENERIC_GISAID_LOG//PLATE/$PLATE}" \
--fastq_dir ${FASTQS} \
--plate "${PLATE}"Python:
ncbi.write_presubmission_metadata()NOTE: Once you're ready, you can drop the --test_mode and --test_dir flags
Shell:
# if submitting to BioSample and SRA (and if creating a new BioProject):
ncbi_submit ftp submit \
--db bs_sra \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--outdir "${NCBI_DIR}" \
--fastq_dir "${FASTQS}"
# if only creating a new BioProject:
ncbi_submit ftp submit \
--db 'bp' \
--plate \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--subdir "${NCBI_SUBDIR}" \
--outdir "${NCBI_DIR}"
# wait a while and try this to download reports and view submission status
ncbi_submit ftp check \
--plate \
--db bs_sra \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--outdir "${NCBI_DIR}" Python:
# if submitting to BioSample and SRA (and if creating a new BioProject):
ncbi.submit(db="bs_sra")
# if only creating a new BioProject:
ncbi.submit(db="bp")
# wait awhile and try this to download reports and view submission status
ncbi.check(db="bs_sra")(NOTE: not fully tested) To link your fasta in GenBank to the associated reads, you'll want to add in the BioSample accessions before submitting.
- Acquire BioSample accessions via one of these methods:
- download accessions.tsv file from NCBI and then use
ncbi_submit- (Do this if you submitted to BioSample via NCBI's Submission Portal)
- use
ncbi_submitfor everything- (Do this to avoid manual uploads via NCBI's Submission Portal) Shell:
- download accessions.tsv file from NCBI and then use
# dowload report.xml files to get accesssions from
ncbi_submit ftp check \
--db ${DB} \
--outdir "${NCBI_DIR}" \
--config "${NCBI_CONFIG}" \
-u "${ncbi_username}" \
-p "${ncbi_password}" \
--plate "${PLATE}" \
--fastq_dir "${FASTQS}"
# add accessions to genbank.tsv
ncbi_submit --prep_genbank \
--outdir "${NCBI_DIR}" \
--config ${NCBI_CONFIG} \
--fasta "${GENERIC_CONSENSUS//PLATE/$PLATE}" \
--plate "${PLATE}"
# submit to GenBank (NOTE: db='gb')
ncbi_submit ftp submit \
--db gb \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--outdir "${NCBI_DIR}" \
--fastq_dir "${FASTQS}"Python:
# dowload report.xml files to get accesssions
ncbi.check(db="bs_sra")
# prepare genbank submission files and submit
ncbi.submit(db="gb")
## or
# files can also be prepared without submitting via:
ncbi.write_genbank_submission_zip()Wait awhile (10+ minutes) for NCBI to start processing the submission. Then run this to download reports and view submission status. This works for whichever db you want to check on. If not specified, you'll get results on all submitted dbs.
Shell:
# check GenBank submission status (NOTE: db='gb')
ncbi_submit ftp check \
--db gb \
--test_mode --test_dir \
--config "${NCBI_CONFIG}" \
--outdir "${NCBI_DIR}"Python:
# check GenBank submission status (NOTE: db='gb')
ncbi.check(db='gb')To acquire the accessions for all samples submitted via ftp under your group's account, ncbi_submit can download all xml report files and parse out the accession details. A directory will be created in outdir containing all submission-specific directories, each containing its report files. The -f or --files flag allows the use of a list of report files to parse. If provided, those files will be parsed for accession details rather than downloading the latest report files. NCBI only stores uploads for a certain amount of time, so accessions found in newly downloaded reports are combined with those from previously downloaded report files to get the most complete picture. This means it's important that you run ncbi_submit ftp check after each submission has been processed to ensure accurate results. The database can be specifed to indicate which accessions are desired and yield csvs (for the BioProject associated with your current config file) at <outdir>/accessions_<bioproject>.csv with the following fields:
| database | fields |
|---|---|
| 'bs_sra' | sample_name, BioSample, SRA |
| 'bs' | sample_name, BioSample |
| 'sra' | sample_name, SRA |
Shell:
ncbi_submit ftp get-accessions \
--db "bs_sra" \
--config "${NCBI_CONFIG}" \
--outdir "${REPORT_DIR}" \
-u "${ncbi_username}" \
-p "${ncbi_password}" \Python:
ncbi.get_all_accessions(db="bs_sra")Shell:
ncbi_submit ftp get-accessions \
--db "bs_sra" \
--config "${NCBI_CONFIG}" \
--outdir "${REPORT_DIR}" \
-u "${ncbi_username}" \
-p "${ncbi_password}" \
-f s1/report.xml s2/report.xmlPython:
ncbi.get_all_accessions(db="bs_sra",report_files=["file1", "file2"])If you want to update the reads for a sample you've already submitted, you must do the followind:
- Email nlm-support@nlm.nih.gov and supply them with a list of SRA runs to suppress.
- Once suppressed, you can upload a new version of the sample where the
submission.xml
- references the BioSample (rather than submitting a new BioSample block) and
- has a new, unique SPUID for the SRA action block.
The submission.xml can be prepared as shown below and then submitted as discussed previously in File Submission. Whereas normally an error would occur if a previously-submitted sample appears in the seq_report file, the flag --update_reads tells ncbi_submit to search for BioSasmple accessions of and include previously-submitted samples in the submission.xml. In most cases, if you are updating reads for a sample, a new SRA spuid is required. The --spuid_endings flag takes a parameter mapping samples that are being updated to a suffix. For any explicitely names samples, the suffix(es) will be added at the end of the automatically-generated SPUID. Usually '2' is a good suffix choice (unless another update has already been made using that same suffix for the sample of interest).
These are not currently supported but could be added in the future if they seem important/useful.
Shell:
ncbi_submit file_prep \
--config "${NCBI_CONFIG}" \
--seq_report "${SEQ_REPORT}" \
--outdir "${NCBI_DIR}" \
--fastq_dir ${FASTQS} \
--plate "${PLATE}" \
--update_reads \
--spuid_endings 'suffix1:samp1,samp2;suffix2:samp3'Python:
ncbi.write_presubmission_metadata(update_reads=True,spuid_endings={"sample1":"suffix1", "sample2":"suffix1", "sample3":"suffix2"})config: Contains preset values and details about your lab, team, and submission plans that are necessary for submission.seq_report: Main metadata file with sample details - can be equivalent to NCBI's BioSample TSV for use with the Submission Portal.
exclude_file: Contains a list of "sample_name"s to exclude from NCBI submission (each one on a new line).barcode_map: Used as a cross-reference. If all samples frombarcode_mapappear inseq_report, that's great. Otherwise, you'll get a warning with directions for adding samples to theexclude_fileif they shouldn't be submitted. File should have no headers. Lines must be: "{barcode}\t{sample_name}".
fastq_dir: Required forfile_prepandftpif submitting reads to SRA. Indicates where the fastqs should be gathered from. Any fastqs with "sample_name" values that aren't supposed to be submitted will be ignored.outdir: Highly recommended but will defualt to "./ncbi" or "./ncbi_test". A directory to house output (submission reports,exclude_file, output fromfile_prep). Will be created, if needed.subdir: Only used forftptasks. A prefix to use for submissions for the given dataset. Defaults toplate, if plate is provided.
| File type | BioProject | BioSample | SRA | GenBank | Description/Link |
|---|---|---|---|---|---|
| Webpage | ✓ | ✓ | ✓ | ✓ | Protocols & TSVs for use at Submission Portal |
| XML | create | create | create | SRA submission w/ new BioSample & BioProject | |
| XML | link | create | create | SRA submission w/ new BioSample & existing BioProject | |
| XML | link | link | create | SRA submission w/ existing BioSample & BioProject | |
| XML | create | GenBank XML | |||
| doc | example | Example GenBank submission zip | |||
| XSD | schema | BioSample XML Schema | |||
| XSD | schema | BioProject XML Schema | |||
| err | validate | Submission Error Explanations |