How Bioinformatics Scientists Use iPhone Notes for Reproducible Analysis

Bioinformatics analyses are scientific decisions as much as computational processes. The cutoffs chosen for variant filtering, the normalization method applied to RNA-seq counts, the reference genome version used — all affect biological conclusions. Scientists who document these decisions create reproducible analyses; those who rely on memory create papers that cannot be reproduced.

Why Bioinformatics Scientists Need Systematic Notes

A bioinformatics scientist may run five to ten active analyses simultaneously across different projects, collaborators, and biological questions. Each has its own dataset, pipeline version, QC thresholds, and biological interpretation context. Without notes, the reasons for non-obvious decisions evaporate between analysis and manuscript.

Pipeline Decision Notes

For every analysis, document the non-default choices:

• Reference genome version — hg38 vs. GRCh38, GRCm39, and why
• Aligner and version — STAR 2.7.x, BWA-MEM2, minimap2 — with citations
• Trimming parameters — adapter sequences, minimum quality, minimum length
• Filtering cutoffs — mapping quality, duplicate marking, depth thresholds
• Normalization method — TPM, RPKM, TMM, DESeq2 normalization — rationale
• Statistical model — DESeq2 vs. edgeR vs. limma, and why for this data type

Pipeline decision notes allow you to exactly reproduce an analysis six months later when the manuscript reviewer asks for it.

QC Decision Notes

Quality control involves judgment calls:

• Samples failed QC — which samples were excluded and specific metrics that triggered exclusion
• Outlier samples — identified by PCA or clustering, and whether excluded or retained
• Batch correction applied — ComBat, limma removeBatchEffect, and the batch variables used
• Sex and ancestry checks — concordance with metadata
• Contamination assessment — any samples with cross-sample contamination evidence

QC decision notes are the audit trail that supports methods section claims about "quality control filtering."

Variant Interpretation Notes

For variant analysis in genomics or population genetics:

• Variant filtration cascade — each filter applied and the fraction of variants retained
• Functional annotation — VEP version, databases used (gnomAD, ClinVar, COSMIC)
• Pathogenicity classification criteria — ACMG criteria applied and their evidence
• Novel variant interpretation — reasoning for pathogenic/benign classification
• Candidate variant prioritization — what made variant X the lead candidate

Variant interpretation notes are especially important for variants of uncertain significance that may be reclassified as databases grow.

Analysis Version Notes

Analyses evolve across a project:

• Analysis version number — v1, v2, v2.1
• What changed from previous version — data update, bug fix, parameter change
• Impact on results — how did results change?
• Who requested the change — collaborator, reviewer, PI

Version notes prevent confusion when a collaborator references "the analysis from the September meeting" that has since been superseded.

Collaboration Notes

Bioinformatics is inherently collaborative:

• Collaborator name and institution
• Data received — what files, when, in what format
• Data transfer method — secure transfer, dbGaP, IRB requirements
• Questions posed by collaborator — what biological question they need answered
• Deliverables and timeline — what analysis outputs are expected and when

Collaboration notes prevent the common confusion about which version of data was analyzed for which version of a manuscript.

Interpretation Notes

Biological interpretation requires separate documentation from computational results:

• Top hits and their biological significance — why gene X is interesting
• Pathway analysis interpretation — which pathways are enriched and what that means in the disease context
• Comparisons to literature — how your findings compare to published datasets
• Alternative interpretations — what other biological explanations fit the data

Interpretation notes feed directly into manuscript discussion sections and grant reports.

FAQ

Q: How do I note changes to analysis when a manuscript is under review? A: Create a revision analysis note: what the reviewer asked for, what analysis was run, and how results differed from the original submission. This keeps revisions organized and reproducible.

Q: What about notes on negative results? A: Negative results are as important as positive ones — note what was tested, what parameters were used, and why the result was negative. This prevents repeating failed analyses.

Q: How do I track database versions I've used? A: A database version log: gnomAD version, ClinVar download date, Ensembl release — update it each time you run a major analysis. Database version affects variant allele frequencies and interpretations.

Q: Should I note when I consult published pipelines vs. develop my own? A: Always — "pipeline adapted from [citation] with following modifications" is a methods statement that requires accurate notes on what you actually did differently.

Q: How do I organize notes for multi-omics projects? A: One note per data type (WGS, RNA-seq, ATAC-seq, proteomics) with its own pipeline decisions, then an integration note documenting how the data types were combined.

Q: Can I use notes to track compute resource usage? A: A resource tracking note per analysis (cluster, CPU hours, memory, storage) supports grant budget justifications and helps you estimate future computational needs.

Related Reading

• How computational biologists use iPhone notes for research
• How researchers use iPhone notes for scientific documentation
• How data scientists use iPhone notes for analysis work
• How clinical research coordinators document study data

Sources

• FAIR data principles and bioinformatics reproducibility guidelines
• Bioconductor project, RNA-seq workflow documentation standards
• Nature Methods, bioinformatics reporting standards and methods checklists