Updating Assemblies

Updates to assemblies can be performed through Webin-CLI, and for all except SARS-CoV-2 assemblies*, result in the creation of a new version of the assembly.

Updates to assemblies are performed in very similar fashion to submission of assemblies. It is therefore sufficient to follow the guidance in the assembly submission pages, while considering the below factors. Note that this information applies only to updating the sequence and annotation, not the general assembly metadata.

*To maximise the speed of processing, sequence updates to SARS-CoV-2 assemblies are no longer versioned. This means that any such update will result in completely new sequence and ERZ assembly accessions. In exceptional cases we may be able to maintain versioning for a SARS-CoV-2 sequence. Please contact us via: virus-dataflow@ebi.ac.uk to discuss this.

Study and sample reference must be maintained:

To submit an assembly update, make sure you reference the same study and sample accessions as were used in the original submission. In fact, this study-sample pair is unique to your assembly and is the means by which you submission is recognised as an update rather than a new assembly.

Chromosome names must be maintained:

If you are updating a chromosome-level assembly, all chromosomes in the original assembly must be present in the updated version, with identical name. Webin-CLI does not currently validate for this and the error will only be caught after submission. If you believe you may have mistakenly submitted an update with a chromosome absent or incorrectly named, please contact our helpdesk.

The chromosome name is the value our pipeline uses to identify which sequence in your update should be considered the successor to each of the original chromosomes. The name is specified in the chromosome list file, as well as in the identifier lines of your FASTA or the ‘AC * ‘lines of your EMBL file.

Chromosomes may be added:

Note that it is acceptable to include new chromosomes, as long as all previous chromosomes are maintained. Also, it is acceptable to update a contig or scaffold-level assembly to a chromosome-level assembly, in which case there are not yet any chromosome names to maintain.

Sequences other than chromosomes may be removed:

Removing sequence is acceptable where the sequences being removed are not chromosomes and the chromosome set is otherwise unchanged. This is the normal route for removing contamination, or for dropping scaffolds and non-chromosomal contigs from a later version.

Removing a chromosome is a different case, and cannot simply be submitted: contact our helpdesk first. We may need to enable the update before you submit it, and to suppress the removed chromosome from the previous version afterwards.

If chromosome names change, keep the object names stable:

Where you are renaming chromosomes, keep the object names in the chromosome list file the same as in the previous version, so that each sequence can still be mapped to the correct existing accession.

Submit the update in the same format as the original:

If the original assembly was submitted as an EMBL flat file, submit the update as a flat file. Where the original was submitted as FASTA, a FASTA update carrying the same metadata is expected.

Annotation status must be maintained:

If your assembly was previously submitted with annotation and you are updating the assembly you can do this following the above factors, i.e. maintaining the study and sample, and adding an updated assembly including annotation.

If your assembly did not originally include annotation, it is not possible to update your assembly with annotation. An unannotated assembly can only be updated with new versions of the same unannotated assembly. In order to add annotation to a previously unannotated assembly you need to re-submit the assembly. Note this will result in new accession numbers. You must first contact the helpdesk to request suppression of the existing assembly version.

Both the sequence and annotation may be updated, or just one:

When updating an annotated assembly, it is not necessary that both the sequence and annotation have been changed. Just one having changed is sufficient. In either case, the assembly version and all sub-sequence accessions will be incremented.

Use a new ASSEMBLYNAME:

The Webin-CLI manifest format includes an ‘ASSEMBLYNAME’ field. This must be unique in each of your submissions, whether they are updates or new assemblies. To check the name used for the previous submission, look up its ERZ accession in interactive Webin. Enter the XML view and find the ‘alias’ field, which will resemble “webin-genome-<ASSEMBLYNAME>”. Using a pre-existing assembly name will result in an error informing you the object you are trying to submit already exists:

ERROR: In analysis, alias:"webin-genome-<ASSEMBLYNAME>", accession:"". The object being added already exists in the submission account with accession: "ERZxxxxxxx". The submission has failed because of a system error.

Taxonomy changes require our assistance:

If there is a problem with the taxonomy of your sample, please contact our helpdesk. This process will require first that the sample taxonomy be correct, so you are welcome to update your sample first. If the taxon you wish to use does not exist in the taxonomy database, you will need to request its addition.

Tip

The project accession and sample accession associated with an assembly cannot be changed after submission. Please make sure you link your assembly to the correct project and sample during submission.

Contact us for other metadata updates:

The advice given here pertains to updating the sequence and annotation (if applicable) of your assembly. If you wish to change the assembly metadata values, please contact our helpdesk with the assembly accession or accessions affected and the exact change required. This route covers values such as the assembly name and description, and sequencing kit information.

It also covers refreshing an assembly after a linked record has changed. If you have corrected the organism name or taxon on the linked sample, tell us that it has been updated, so that the relevant pipeline can be run and the associated assembly or WGS record refreshed to match. Allow a few days after reprocessing for the change to be reflected publicly.

State of publicity will be the same:

If the original assembly is already public, then the update will be made public as soon as processing is complete. However, please be aware that this process can take some time to complete, and you should allow at least a week after successful submission for your update to be apparent.

Allow 24 hours after original submission:

Please submit any update at least 24 hours after the original submission, to avoid any issues in processing and delays in publication.

Metagenomic Assemblies:

Primary and binned metagenome assemblies cannot be updated. If you wish to update one of these assemblies, please contact our helpdesk so that we can suppress the existing version. However, MAG assemblies work the same as isolate assemblies in the context of updates.