Computing the checksum locally and comparing to the published value
is the standard verification step. Commands per OS:

Linux / macOS terminal:

# MD5
md5sum file.zip                    # Linux
md5 file.zip                       # macOS

# SHA256
sha256sum file.zip                 # Linux
shasum -a 256 file.zip             # macOS

Windows PowerShell:

Get-FileHash file.zip -Algorithm MD5
Get-FileHash file.zip -Algorithm SHA256

Compare the output (a long hex string) to the value on the download
page. If they match exactly, the file is intact. If they differ,
re-download — the difference is not normally something you can fix
after the fact.

Two paths depending on where the Docker image lives:

From a container registry (Docker Hub, quay.io, GHCR):

singularity pull docker://quay.io/h3abionet_org/imputation_tools:latest
# produces imputation_tools_latest.sif

From a local Docker daemon (docker images shows it):

docker save myimage:latest -o myimage.tar
singularity build myimage.sif docker-archive://myimage.tar

Common issues:

• ENTRYPOINT differs from Singularity behaviour — singularity exec invokes a command directly; singularity run runs the
  runscript. If your Docker image relies on ENTRYPOINT, you may
  need to invoke singularity run, or pass the original entrypoint
  explicitly.
• User-id remapping — Docker runs as root by default;
  Singularity runs as your user. Files written to mounted volumes
  will be owned by you, not root.
• Bind paths — pass --bind /scratch:/scratch (or similar) to
  make HPC scratch directories available inside the container.

For an AfriGen-D / H3ABioNet pipeline, the recommended approach is
to use the Nextflow singularity profile, which handles the pull
and caching automatically.

Several catalogues index African genomic resources:

• H3Africa Data & Biospecimens Catalogue —
  https://catalog.h3africa.org. Searchable index of H3Africa
  consortium datasets and biospecimen collections, with contact
  information for each study.
• AfriGen-D Reference Panels Explorer —
  https://fedimpute.afrigen-d.org (Explorer tab). Lists the imputation
  reference panels available on the AfriGen-D service, with sample
  counts and per-population allele frequency information.
• eLwazi Open Data Science Platform —
  https://elwazi.org. The platform aggregating African data
  science resources, including pointers to consortium catalogues.

Most H3Africa datasets are controlled access — discovery is open,
but obtaining the data requires a Data Access Agreement (DAA) with
the originating study. The catalogue entry for each dataset lists
the contact / data access committee responsible.

For dataset-specific questions that the catalogue can't answer
(custom subsets, joint analyses), reach out to the AfriGen-D /
H3ABioNet leadership directly via the helpdesk — we'll route to
the right collaborator.

The standard converter is vcf2maf from Memorial Sloan
Kettering Cancer Center:

# Clone the repository
git clone https://github.com/mskcc/vcf2maf.git
cd vcf2maf

# Install Perl dependencies
cpanm --installdeps .

# Convert
perl vcf2maf.pl --input-vcf sample.vcf --output-maf sample.maf \
  --tumor-id SAMPLE_TUMOR --normal-id SAMPLE_NORMAL \
  --vep-path /path/to/vep --vep-data /path/to/vep_cache \
  --ref-fasta /path/to/GRCh38.fa

vcf2maf requires Ensembl VEP for variant annotation, so install
VEP first if you don't have it. The tool annotates each variant
and selects the canonical transcript effect, which is what MAF
format expects.

For a quick conversion without full annotation, maftools::vcfToMaf
in R can take a vcf2maf-style TSV or a pre-annotated VCF and
produce a maftools-ready data frame — useful if you only need
the conversion to drive maftools visualisations and not the full
cancer-genomics annotation set.

Alternatives: bcftools + a custom awk script can produce a minimal
MAF for simple cases, but vcf2maf is the standard most maftools
workflows expect.

Open a helpdesk ticket at https://helpdesk.afrigen-d.org/helpdesk/
and choose the queue that best matches:

• Genotyping & Imputation Service — anything to do with
  imputation jobs, reference panels, or fedimpute.afrigen-d.org.
• Pangenome & Reference Graph — questions about the pangenome
  reference work.
• System/Technical Administration — server access, account
  issues, infrastructure.
• General Enquiries — anything that doesn't fit a specific queue.

To help the team respond faster, include:

1. What you were doing — the URL or feature you used.
2. What you expected — the result you wanted.
3. What happened instead — the error message (paste full text;
   screenshots are also useful).
4. The Job ID or ticket reference — if the issue is about a
   specific imputation run or earlier ticket.

You'll receive an email confirmation with the ticket number.
Replies to that email automatically add to the ticket thread, so
you can continue the conversation by replying to email.