What makes a genome report traceable

A polished PDF does not by itself show that a result is reliable. Readers need to know which file was analysed, which reference and method were used, and what limitations apply to that run. This article follows the information that matters from raw data to report.

Start with the input

FASTQ files contain sequencing reads and a quality value for each base. In paired-end data, R1 and R2 belong together. A missing file or mismatched pair can affect the steps that follow.

BAM and CRAM files contain reads that have already been aligned. They usually need an index for quick access to individual regions; CRAM also depends on the matching reference. A filename or extension alone cannot confirm that the contents, index and reference are compatible. A workflow should check the expected headers and companion files.

Genome processes selected genomic files on the Mac. Each analysis may require additional tools, indexes or reference data. A report should identify the actual inputs and workflow rather than infer its data basis from a filename. The guide to searching for rsIDs directly in a BAM file shows one practical use case.

Reference and coverage are part of the result

Reference assemblies such as GRCh37 and GRCh38 use different coordinates. Even the same assembly label does not necessarily describe every detail of a reference file. Coordinates are meaningful only alongside the reference used.

An emitted variant call also does not show how well reads support that position. Coverage, allele support and quality values help assess the technical evidence. An uncallable or missing genotype does not prove that a variant is absent.

Different questions require different methods

Genomic analysis is not one universal test. HLA typing, pharmacogenetic analysis, mitochondrial variant analysis and polygenic scoring use different data, tools and rules. Each result addresses a limited technical or scientific question.

An HLA marker is not a diagnosis. A pharmacogenetic phenotype does not automatically determine a medicine dose. A polygenic score is not an individual percentage without a suitable comparison model. Reports should state the method, source and known limitations alongside the relevant result.

What a traceable report records

Depending on the workflow and report, this can include the input files, reference assembly, tool and database versions, relevant parameters, quality information and sources. Not every step produces the same metadata. Missing information should remain identifiable instead of being replaced by an assumption.

If a report lists checksums such as SHA-256, they can help establish whether a file’s bytes have changed. A checksum does not prove that the file is scientifically correct or that a later run will produce the same result. Traceability makes a result easier to review; it does not guarantee reproducibility or freedom from error.

PDF is a useful fixed format for reading and sharing a report. It can keep text, tables, sources and page references together. Quality still depends on correctly assembled content, clear limitations and a clean export. Creating a PDF does not make an analysis scientifically reliable by itself.

Files stay local; selected lookups go online

Genome processes genomic files locally and does not upload them to PJ Labs. Some features can still use the network: an optional reference lookup sends the requested rsIDs to the relevant public database service. The genomic dataset itself is not sent for that lookup. Keep these separate data flows in mind when using the software.

Genome is intended for research, information and education and is labelled Research Use Only. A technical report does not replace medical advice, diagnosis or treatment. For further context, see How to read a genome report and the Genome help.