So I should re-send my original emails? ok, thanks.
• 0 views
•
link
written
by
susan.klein •
NCBI Genomes submission - what to do if FCS report is incorrect? Stuck with nowhere to go!! No communication!
Hi,
is anyone else having problems with communication with NCBI for submissions? I have found previously that replies can take over a year.. and in the current situation I expect that it is now worse.
I have two plant genomes (from many others that I have fixed) that cannot pass their FCS error report. One says it is too large - but the existing RefSeq Genome for the species is probably incorrect - it is old Illumina assembly and is missing very large repeat regions, my data is Pacbio HiFi and the assembly checks out against other species in the group.
The other genome has a 'contamination' that is a Rhizobium transposase gene - this is almost certainly horizontal gene transfer, so it needs to be left in the contig (about 3kb in a 16Mbp contig).
The genomes submission portal only allows 30 days to get the submission fixed then it is automatically deleted. I have emailed NCBI but I think it is very unliekly they will reply in 30 days, if at all..
Is anyone else having these sserious issues? I find it puzzling that there does not seem to be more fuss about it online?
I have also tried ENA but hit roadblocks with their Aspera connection constantly being dropped, and their genomes submission requiring reads, which for some of my genomes are files >100Gb, so very unlikely to ever transfer succesfully. Similar to NCBI, they also never reply to emails.
Anyone have any suggestions?
Thanks.. getting desperate!
S.
• 2,083 views
•
link
updated
by
MirianT_NCBI •
•
written
by
susan.klein •
•
written
by
susan.klein •
3 answers
Hi,
I'm so sorry about this situation. I'm not involved in submissions at NCBI, but I reached out to the submissions team to inquire about your situation.
They asked you to please write to them again at genomes@ncbi.nlm.nih.gov and provide the SUB number for the submissions in question and the evidence to support that the contamination calls are invalid. They will look into it and review the evidence.
I hope this helps.
Mirian
• 0 views
•
link
written
by
MirianT_NCBI •
I have resent them . SUB15942405. One FCS failure is a horizontal transfer, and the other one is an expected size failure based on an out of date and old Illumina assembly which is lacking large portions of repeats. Thanks.
• 0 views
•
link
written
by
susan.klein •
I got replies from NCBI and the size failure genome was accepted. Only one more that had horizontal transfer mistaken as contamination remains. NCBI said it was sent to FCS expert for evaluation. No update yet.
Thanks for the help!
• 0 views
•
link
written
by
susan.klein •
All genomes now accepted. Very happy, just need to get the last 6 SRA files to upload..
• 0 views
•
link
written
by
susan.klein •
I am so glad to hear that everything got sorted out! :)
• 0 views
•
link
written
by
MirianT_NCBI •
So my question remains how to people get past this step if FCS is incorrect?
I had to demonstrate that their claims were wrong.
I recently had to prove that NCBI's own records were the source of the error. One of my genomes was flagged as contaminated because the majority of its sequences aligned with a reference genome from a completely different genus. To resolve this, I had to demonstrate that the reference genome itself was misclassified and that this misclassification wasn't just affecting my submission, but it was causing the systematic suppression of my genome and several others of the same species.
After an exchange of ten emails, they finally unsuppressed my genome along with hundreds of others.
In your case, you must demonstrate to them that:
- The transposase gene in question is a legitimate horizontal gene transfer (HGT) event.
- The existing RefSeq genome for the species is likely incorrect or incomplete because it is an older Illumina assembly that lacks many large repeat regions.
• 0 views
•
link
written
by
andres.firrincieli
Thanks, it's good to get some idea of what happens. I'll just have to hope they reply within 30 days.
• 0 views
•
link
written
by
susan.klein •
Eventually all the issues were resolved. NCBI accepted supressed the erroneous FCS reports and accepted all 29 genomes, just withing the deadline, so all good. The SRA uploads were also resolved although it took many attempts to get all the files uploaded.
• 0 views
•
link
written
by
susan.klein •
I'm so glad to hear that everything got resolved within the deadline!
• 0 views
•
link
written
by
MirianT_NCBI •
Log in to answer this question.
More posts like this
-
NCBI table2asn has "SeqGapProblem" error with JBAT-assembled genome
written by Arielle •
-
How to check qulaity of genomes present at NCBI?
written by genious.kalia.141198 •
-
Submit paired bacterial chromosome and plasmid scaffolds to NCBI?
written by causland •
-
Eukaryotic gene annotation transfer tools
written by erik.kastman •
-
NCBI: batch upload WGS sequences linked to bioproject/biosample
written by ALchEmiXt
-
Genbank to bed conversion for bedtools analysis
written by susan.klein •
-
MG-RAST upload problem
written by susan.klein •
-
Preparing A Scaffolded Genome For Genbank Submission
written by Rvosa
-
Downloading Ncbi Query Results Using Aspera?
written by edwardbeaty •
-
Problems Running The Ncbi-Blast Tools
written by Naren
In my experience the NCBI staff always respond, but they are not prompt. I think for this purpose they often have to hire a contractor. There is not much back-and-forth with them, which is why the whole process is largely automated. That may not be to your liking, and you can always try to send them pre-emptive messages always citing the ID of your submission. They may or may not read them, however.
My suggestion is to move on with next steps and don't hold your submission or publication to get this done. Ideally, one starts these things well before the manuscript is ready to go out, and then they can hold the results until after the publication if appropriate.
I am sure many people are having problems. Still, it is very likely that they get thousands of submissions per day, and this will never be an efficient process until they hire hundreds of curators. I suspect they have at least an order of magnitude fewer people than that.
Hi, thanks for that but what does "My suggestion is to move on with next steps" mean? There are actually no further next steps - I am not holding a publication (although that might be a problem in near future), In the submission process, when the FCS identifies a genome needs fixing, you must fix it to move on (the next step, as it were is NCBI accepting the submission and issuing an accession). If you don't fix it - as in this case because the FCS is incorrect, there is nothing you can do until they reply, and if that is not within 30 days, then it is deleted and you have to start again - this is catch 22, there is no way to move forward. So my question remains how to people get past this step if FCS is incorrect? For example, a Google search reveals nothing on this. There is literally no way to get my genomes submitted! I hope you understand? I suppose my hope is that there is someone from NCBI who reads this forum. Thanks, S.
Still no reply from NCBI, in case anyone was wondering. Also having trouble uploading reads to SRA now. Using aspera on command line, it will usually upload one file (of 52 total) then NCBI closes the connection.. also there is no way to check which files have succeeded, except by submitting the folder again and getting a missing files report... may have to do this 52 times.. but might be failing compeltely, also can't be sure the files are completely uploaded, the values in aspera log certainly don't match. The files are large - 20 - 150 Gb.
I hear you that this is all annoying, but only NCBI can help with that, and with all the shutdown situations, reductions in funding of the NIH, and uncertainties in the US right now, why not just uploading it all to Zenodo? Inform the journal editor about the ingoing trouble with you submit your maniscript and call it a day.
Have you tried these uploads from a different location/ISP? It is certainly possible that local firewall/deep packet inspection device could be causing the uploads to fail.
If there was a generalized issue with SRA uploads then there would have been some sort of an announcement.
I can only upload from the HPC that I use. They have assured me that they have no firewall for this connection. It is NCBI closing the connection according to the error messages, e.g.:
Session Stop (Error: Mandatory management socket localhost:42779, Connection reset by peer (server))
These kind of network errors can be difficult to diagnose since there is no way for us to understand the exact networking setup at your institution. It is minimally likely that there is a proxy involved (HPC is unlikely to have a direct internet connection) and that may be the cause of the problems.
Can you take the smallest file that you have and try uploading from a different location. If that goes smoothly we will know where the issue lies.
This is definitely an NCBI connection issue. I have 10 files left now, apsera is uploading one or two every 10 hours (the walltime for my HPC jobs). So I will continue running it until they are done. If the files are not complete (no way to know yet), then I'll give up and use Zenodo as you suggest.