Update diversityOfIlealMucosa - user feedback

[arschat]

In the hca-support channel (i.e. contact us form from Data Portal) we received the following message:

Check out this new ticket w/ subject: SRR to SRX mismatch.
Ticket description:
----------------------------------------------
in the metadata file diversityOfIlealMucosa 2024-11-28 00.27 SRR19039103 is listed as SRX15110551 but on ncbi SRA it's https://www.ncbi.nlm.nih.gov/sra/?term=SRX15110547.
Similarly SRR19039104 on SRA is https://www.ncbi.nlm.nih.gov/sra/SRX15110546 not 551 like the metadata here.
Are these metadata supplied by the authors or converted from SRA or somewhere else?
------------------
Submitted from: https://explore.data.humancellatlas.org/projects/9dd91b6e-7c62-49d3-a3d4-74f603deffdb/project-metadata.

[arschat]

Using the SraRunTable.csv, I checked in metadata spreadsheet the following (comparison spreadsheet: diversityOfIlealMucosa_metadata_23-08-2024_check_map.xlsx):

• Donor level, biosample id with
  • age
  • ethnicity
  • sex
  • disease
• specimen level, based on biosample id
  • disease
  • srx id
  • srs id
• cell suspension, based on srx id
  • samn id
  • gsm id
• sequence file, based on srr
  • input biomaterial (srx id)
  • insdc srx id
  • srr in filename
• analysis file, based on gsm from filename
  • input biomaterial (srx id)

Discrepancies found

• There are 2 columns with SRX mis-match on sequence file tab.

  • input biomaterial (srx id) cell_suspension.biomaterial_core.biomaterial_id: 30 values are identical
  • insdc experiment (srx id) process.insdc_experiment.insdc_experiment_accession: 121 values are missing or non valid

• There are 2 runs that don't have sequence files: SRR19039060, SRR19039068

  • corresponding srx ID not missing from cell suspension nor analysis file

• The filenames does not seem to match the contributor provided nor the archive generated files

  • i.e. for SRR19039056 the contributor provided is named p21063-s007_GCA29_5GEX_E2_S13_L003_I1_001.fastq.gz while in the spreadsheet the name is SRR19039056_1.fastq.gz
  • number of files for archive generated files is 55 while in spreadsheet we have 127 files. the contributor provided files are 133 which align with the two runs missing (133 - 3*2 = 127 files)

• hca-util upload area 46f49251-6bd0-4d48-8e00-bdbcd5864072

[arschat]

All srx mismatches have been fixed and submission is now in graph valid.

• ncbi cloud delivery requested for two runs
• files delivered
• files renamed & moved to hca-util-upload-area
• files added to submission
• patch content for each file
• create 2 processes and link all 6 files
  • SRX15110562 -> SRR19039060
  • SRX15110554 -> SRR19039068
• replace old file schema & update ontology fields data -> EDAM
• submission export
• import form sent

[arschat]

File GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz is invalid in archive (downloaded twice and got the same error)

fastq_validator.sh GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz
Checking integrity of gzip file GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz...done.
Checking  GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz
fastq_utils 0.24.1
DEFAULT_HASHSIZE=39000001
Scanning and indexing all reads from GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz
CASAVA=1.8
57300000
ERROR: Error in file GCA9-22812-5GEXC8_S3_L001_R2_001.fastq.gz: line 229267721: read length too small - 0

Will include "archive generated file" for this run.

[arschat]

import form sent