Blog

A practical comparison of whole-genome and whole-exome sequencing, with the pipeline, tools, flags, and failure modes we would use to analyze either one.

A working pipeline that classifies the variants in a personal VCF under ACMG/AMP rules, and why automated calls disagree with expert curation.

A blood panel measures a few dozen analytes in one tube at one moment. Here is what that covers, why single results mislead, and how to store and analyze your own results over time.

The reagent cost of a 30x human genome is now in the low hundreds of dollars, but the price you pay depends on coverage, data delivery, and whether the lab is CLIA-certified. Here is the full cost stack, including storage and compute.

A step-by-step guide from raw FASTQ to an annotated variant table on your own machine, covering alignment, calling, QC thresholds and annotation.

A working list of blood, molecular, and continuous biomarkers worth tracking on yourself, with the biological-variation math that tells you when a change is real and when it is noise.

How 21 CFR parts 812, 50 and 56 govern clinical investigations of medical devices, including when an IDE application must be submitted to FDA.

Which vitamin and mineral analytes are worth measuring, which ones are nearly uninterpretable in serum, and how preanalytical handling and inflammation decide whether your panel means anything.

A working guide to blood biomarker analysis for technically fluent readers: what a single result can and cannot tell you, how to separate biological variation from signal, and how to structure your own longitudinal data.

A working method for reading your own labs: get the data as structured records, compare against your own history rather than the population reference interval, and use the reference change value to tell a real shift from noise.

The statutory clocks for a security assessment, a standard contract filing and certification of data exports from China, and what extends each.

How to tell Ensembl the database from ensemble methods in bioinformatics, then use both on your own whole-genome VCF: offline VEP annotation, plugin scores, and consensus calling across pipelines.

The three kinds of genetic testing for disease, which platform detects what, and how to work with the raw files yourself.

A working GATK pipeline for calling SNPs and indels from RNA-seq reads, with filters suited to transcriptomic data and a check against genome calls.

How exome capture works, the analysis pipeline for your own FASTQs, and the failure modes that decide whether an exome is the right assay.

How Annex VIII of Regulation (EU) 2017/745 assigns medical devices to class I, IIa, IIb or III, and the rules that decide which class applies.

A step-by-step guide to taking raw whole-genome FASTQs through alignment, variant calling, validation, and annotation, with the specific tools, flags, and quality thresholds we use.

A working end-to-end path from raw FASTQ to an annotated, QC'd variant set using GATK, DeepVariant, and VEP, with the flags, file sizes, runtimes, and failure modes that matter.

Pharmacogenomic testing resolves star alleles in about a dozen drug-metabolism genes and maps them to CPIC phenotypes. Here is what panels cover, why CYP2D6 is the hard part, and how to generate diplotype calls from your own sequencing data.

A SNP microarray genotypes a fixed panel of several hundred thousand predefined positions and, from probe intensities, infers copy number and runs of homozygosity. Here is the data it produces, the tools to process it, and the variant classes it cannot see.