Known Issues
The issues below apply to the CDRv9 release multi-omic data (arrays, srWGS, srWGS SVs, lrWGS, RNA sequencing, proteomics and auxiliary data). We have provided suggested actions for researchers to workaround the issues and provided remediation plans when necessary. Sample lists relevant to these issues can be found on the Researcher Workbench, locations are in the Data Dictionary.
Known Issue #1: srWGS samples were affected by a data quality issue (N=152)
We have identified a data quality issue affecting 152 CDRv9 samples. These samples did not pass the srWGS coverage metrics mean coverage threshold of ≥30x and were included in the callset in order to retain samples with matching multiomics data (RNA sequencing, proteomics, and/or lrWGS). The minimum mean coverage retained was 28.5x. These samples will be retained in future releases. Note that per the selection criteria described for the Long Read samples, these samples had passed srWGS at the time of selection. Coverage fell below the threshold upon reprocessing of srWGS on DRAGEN version 3.7.8 where previously processed and released on DRAGEN version 3.4.12 in CDRv7.
Affects:
- srWGS SNP & Indel samples.
Suggested action:
- We do not believe that this will have an appreciable effect on most downstream analyses, since the mean coverage was close to the cutoff (min: 28.5x vs threshold of 30x). If you believe that your analysis is affected and you are not using matching multiomic data types, remove the samples from your analysis. We provide a list file of research IDs of affected samples in the CDR, see the Data Dictionary.
Remediation:
- These samples will likely be retained in future releases and continued to be documented.
Known Issue #2: srWGS variant sites (0.015%) missing for 8000 samples in known genomic regions
We have identified two regions of missing variant sites (0.015% of the genome) missing on chromosomes 4 and 19 for 8000 participants (1.49% of participants with srWGS data). This was caused by a transient processing issue. The affected regions are chromosome 4: 56,580,000 - 57,040,000 and chromosome 19: 40,097,000 - 40,650,000. Each of the two regions has 4000 participants affected. For each participant, this is 0.015% of their genomic variants.
We will provide lists of research IDs for each chromosome (chromosome 4 and chromosome 19). Each list contains the participants who have the missing genomic variants in that region. Additionally, we are providing two bed files for each missing region and a list of genes that intersect with these regions. See the Data Dictionary for these lists.
Cohort extractions, VCFs, and PGENs are unaffected by this issue.
Affects:
- srWGS SNP & Indel VDS.
- srWGS SNP & Indel VAT.
- srWGS smaller callsets (ACAF, ClinVar, and exome), in the Hail Multi MT, Hail split MT, BGEN, and PLINK bed.
- This issue does not affect VCFs and PGENs.
- Multiomics eQTL, sQTL, and pQTL.
Suggested action:
- We recommend that if researchers are interested in these regions, they remove the affected participants from their analysis.
Remediation:
- We will remediate this issue in the next dataset release.
Known Issue #3: Small amount of multi-omic data missing matching genomic data (N=25)
We have identified research IDs in the srWGS SV, lrWGS, RNA-seq, and proteomics datasets that do not have matching array or WGS data. Please see a list of research IDs of the affected samples in the CDR.
Affects:
- srWGS SV data
- 2 participants have no matching array data, 3 participants have no matching srWGS SNP/Indel data (<0.01% of srWGS SV data).
- lrWGS data
- 22 participants have no matching array data, 23 participants have no matching srWGS SNP/Indel data (0.16% of lrWGS data).
- Multi-omic data
- 1 participant with proteomic and RNA-seq data has no matching array or srWGS SNP/Indel data (0.01% of proteomics data).
Suggested action:
- If matched srWGS or array data is required in your multi-omic data analysis, remove the multi-omic samples from your analysis. We provide a list file of research IDs of affected samples in the CDR, see the Data Dictionary.
Remediation:
- The cause is under investigation. It is possible the srWGS and/or arrays will be restored in future releases. If not, the multi-omic samples will likely be retained in future releases and continued to be documented.
Known Issue #4: Small differences in genetic ancestry categories between multi-omic datasets
Due to a sample processing issue, some samples with multi-omic and/or srWGS SV data will have different genetic ancestry categorizations in these datasets compared to the srWGS SNV/indel data. All differences are to or from the “Remaining individuals” category and affects samples where the confidence was close to the 0.75 threshold (see Appendix G Methods).
For the multiomic data, this impacts 25 samples (0.28%) of RNA-seq and proteomics samples. The differences between srWGS genetic ancestry classifications and the genetic ancestry in pQTLs and eQTLs is expected. However, we believe the differences do not appreciably affect analyses.
For the srWGS SV data, the variant AF fields were calculated using these different genetic ancestry categorizations, which affected 523 (0.5%) srWGS SV samples. These samples were close to the threshold and we believe that these differences are jitter from a second training of the random forest classifier. We believe the differences do not appreciably affect analyses.
Affects:
- The per-ancestry protein quantitative trait loci (pQTL) data and expression quantitative trait loci (eQTL) data of the RNASeq and proteomics datasets.
- srWGS SV VCF fields that use genetic ancestry categories (eg, the population AF fields).
Suggested action:
- We do not believe that any action is necessary.
Remediation:
- The sample processing issue has been resolved and will be remediated for the next dataset release.
Known issue #5: srWGS SNP & Indel variant calls on chromosome Y need additional filtering
We see variants with heterozygous calls in chromosome Y, which cannot be correct germline calls. After manual review, we believe that regions of chromosome Y are prone to misalignment artifacts (low mappability). This will cause heterozygous calls in chrY that are likely artifacts. We have not investigated whether these are somatic mutations.
Affects:
- srWGS SNP & Indel variants: VDS, VCF, Hail MT formats.
Suggested Action:
- If you do not use variant calls on chrY, then take no action.
- Filter the chrY variants using publicly available bed files, such as those published by Genomes-in-a-Bottle.
- Otherwise, we recommend that you use AD, GQ, and GT to filter variants on chromosome Y.
Remediation:
- We will address this in a future release.
Known issue #6: Mitochondrial copy number demonstrates a batch effect due to DNA extraction methods
The srWGS mitochondrial copy number demonstrates a biphasic copy number distribution caused by varying DNA extraction methods. To address this issue, we suggest that researchers include the DNA extraction method (Autogen or Chemogen) as a factor in their analysis. We provide the extraction method for all srWGS samples in a tsv file.
The path for the file is available in the Data Dictionary. For more details about the mitochondrial variant data and methods, see the article How the Genomic and Multi-omic data are organized.
Affects:
- srWGS mitochondrial copy number.
Suggested Action:
- If you do not use mitochondrial copy number data, then take no action.
- Adjust downstream analyses of mitochondrial copy number for DNA extraction method using provided tsv file, available on the Researcher Workbench.
Remediation:
- We will continue to document this effect in future releases.
[Solved] Known issue #7: Incomplete MANE annotations in VAT
Solved August 14, 2026.
Due to a workflow processing issue, the Variant Annotation Table (VAT) was missing Matched Annotation from the NCBI and EMBL-EBI (MANE) annotations. Specifically, only ~6% of the transcripts that should have received MANE annotations were annotated. This was remediated August 14, 2026.
Affects:
- CDRv9 Variant Annotation Table (VAT). The only affected columns were mane_select_name and mane_plus_clinical_name.
Suggested Action:
- If you do not use the MANE annotations, then take no action.
- Otherwise, we recommend that you use the new version of the VAT.
Remediation:
- A corrected dataset is now available on the Researcher Workbench, as of August 14, 2026. The path to this dataset is available within the Data Dictionary.
Known issue #8: Samples were impacted by a data quality issue (N=7)
Reported August 14, 2026.
Seven participant IDs in the srWGS CDRv9 data release are affected by a data quality issue and each of these participants have srWGS and array data. A file containing the affected research IDs is available in the Researcher Workbench (RW). srWGS SV, long-read, RNA Sequencing, and Proteomics samples are not impacted.
Affects:
- srWGS data: all data types and downstream auxiliary data.
- Array data: all data types.
Suggested Action:
- Remove affected samples from analysis. We provide a list of research IDs of affected samples in the CDR, see the Data Dictionary.
Remediation:
- The samples will be removed in the next major dataset release (CDRv10).
Known issue #9: Proteomics samples impacted by a data quality issue (N=165)
Reported October 16, 2026.
We have identified a data quality issue for 165 participant IDs affecting their proteomics data, linked to one Olink Explore HT plate (plate 000012771128; PlateIDs 000012771128_A1_04-09-2025_11-30-10 and 000012771128_A2_04-09-2025_11-30-10; file AoU_Proteomics_LCSET-32810_Extended_NPX_2025-04-29). Arrays, srWGS, SVs, long-read, and RNA Sequencing data are not affected.
We are still investigating the effects on the pQTL and fine-mapping results.
Affects:
- Proteomics (Olink Explore HT) samples on plate 000012771128 (N=165).
- pQTL analyses that include these samples.
- SuSiE fine-mapping results derived from those QTL analyses.
Suggested Action:
- Exclude the affected samples from all analyses using CDRv9 proteomics data. We provide a list file of research IDs of affected samples in the CDR, see the Data Dictionary. If you have already included this plate in an analysis, rerun it without these samples.
- Treat the CDRv9 pQTL and SuSiE fine-mapping results as provisional until we complete our investigation.
Remediation:
- We are still investigating the remediation for this issue and expect to provide an update by November 2026.
Comments
0 comments
Please sign in to leave a comment.