A practical workflow for taking your RNA-seq FASTQ files to transcript quantification, then using web-based analysis tools for the parts they do well. Includes commands, parameters, cost estimates, and the failure modes that waste the most time.
A working pipeline for taking whole-blood RNA-seq FASTQ files to a gene-level count matrix, differential expression across timepoints, pathway scores, and cell-type composition, with the parameters and failure modes that matter.
A working guide to turning RNA-seq FASTQs into the handful of plots that matter: sample PCA, distance heatmaps, MA and volcano plots, and longitudinal gene trajectories, with the commands, parameters, and failure modes.
A practical guide to picking a human genome build and annotation for RNA-seq, building STAR and salmon indexes correctly, and checking that the reference you built is the one your quantification assumes.
A working pipeline for taking a personal WGS VCF into R: GDS conversion, quality metrics, ancestry PCA against 1000 Genomes, variant annotation, and polygenic score computation, with the failure modes that silently corrupt each step.
A working guide to calling, normalizing, annotating, and querying SNPs from your own sequencing or array data, with the specific tools, flags, and failure modes that matter.
A practical walkthrough for taking a whole-genome VCF, keeping only the variants your caller flagged as PASS, normalizing them, and flattening the result into a table you can query. Includes the failure modes that silently drop real variants.
A working list of blood biomarkers that change decisions, why panel size is a poor proxy for value, and how to handle biological variation so you don't chase noise.
Most online biological age calculators are quizzes. Here is what the published estimators take as input, how to compute them from your own blood panel or methylation array, and how much weight the output deserves.
You can draw blood on any cycle day. The question is which analytes move with cycle phase, how much, and how to record the phase so your own longitudinal data stays interpretable.
A practical walkthrough of differential expression analysis in R, from transcript quantification through DESeq2 and limma models, with the design decisions and failure modes that determine whether the gene list means anything.
A practical guide to ordering whole-genome sequencing for yourself: what coverage and read length to insist on, which files to demand, how to check quality yourself, and what the data can and cannot tell you.
Levine's PhenoAge is a nine-marker blood algorithm you can compute yourself in about twenty lines of Python. Here is the formula, the unit conversions that break it, and how it compares to the methylation clocks from the same lab.
A plain answer to what a genome is, what whole-genome sequencing produces, and how to work with your own FASTQ, CRAM, and VCF files without fooling yourself.
A concrete stack for working with personal genome, transcriptome, proteome, and biomarker data: file formats, tools, commands, and where each approach breaks down.
A practical guide to computing principal components from RNA-seq counts in R: which transform to use, how many genes to feed in, how to figure out what each component means, and when PCA is the wrong tool.
A working pipeline from salmon quantification files to a differential expression table and pathway results, written for one person's longitudinal blood transcriptome rather than a two-group lab experiment.
A step-by-step guide to quantifying, normalizing, modeling, and interpreting your own RNA-seq data in R, from FASTQ files through differential expression, pathway scoring, and blood cell deconvolution.
The 56-biomarker performance panels sold as a full-body MOT test the wrong thing in the wrong way. Here is the panel we would order, how to control preanalytics, and how many repeats you need before a number means anything.
A working pipeline for annotating a personal whole-genome VCF: normalize with bcftools, annotate with Ensembl VEP offline, join custom datasets with vcfanno or slivar, then filter down to something you can read.