01 / Data
Three measurement platforms, one statistical treatment
The assay decides how features are counted. Once they are counted it is largely the same problem — protein, spatial and gene-level data all reduce to a feature-by-sample matrix, and only ATAC's genomic intervals need separate handling. That is why one person covers this range rather than three.
The deepest area
- Transcriptomics
- Bulk and single-cell RNA, whole cell or nuclei.
- Chromatin
- Bulk ATAC, single-nuclei ATAC, and paired multiome.
- Genomes
- Exome, whole genome, targeted panels, structural rearrangement.
- Populations
- Genome-wide SNPs, microsatellites, population structure.
- Communities
- Metagenomics, microbiome and antimicrobial resistome.
- Repertoire
- T-cell receptor sequencing, small RNA.
Abundance without counts
- Proteomics
- Differential abundance, normalisation, and missingness treated as a property of the instrument.
- Metabolomics
- The same framework over a different feature space.
- Multi-omic integration
- Several platforms on one cohort, analysed together rather than side by side.
Spatial and in-situ
- Xenium
- In-situ transcriptomics at single-cell resolution.
- Atera
- 10x's newer in-situ platform, analysed on the same footing.
Independent of platform
- Design and power
- What a study can detect, decided before it runs rather than defended afterwards.
- Modelling
- Mixed models, repeated measures, and the choice of replication unit — often what decides whether a result survives review.
02 / Systems
Statistics does not care what the organism is
Which has been fortunate, because the questions have arrived from almost everywhere. Every organism below is named in the bibliography — 87 of them, across 23 taxonomic classes and all four domains of life, from human trauma patients to hot-spring archaea. The first publication was a logistic regression for zoo neonate management in 1995, and the range has widened every year since.
Two papers are worth singling out for scale: five hundred microsatellite loci developed for Peromyscus, and a marker-development method validated across thirty-two species at once. Where a paper covered a group rather than a species — soil communities, the beef resistome — it is counted once, so the true species total is higher than the number above.
03 / Fields
27 fields of study, and counting
The constant is the statistics. The biology has been whatever walked through the door.
Biomedical
- Trauma and transfusion medicine
- Oncology, adult and paediatric
- Cardiology
- Craniofacial development
- Immunology
- Diabetes and metabolism
- Ophthalmology
- Nephrology
- Neuroscience and ageing
- Dermatology
- Rheumatology
- Haematology
- Reproductive medicine
- Obesity and exercise physiology
- Virology
- Infectious disease
- Toxicology
Environmental and agricultural
- Conservation genetics
- Wildlife management
- Fisheries and aquaculture
- Soil and microbial ecology
- Food safety
- Antimicrobial resistance
- Plant science
- Entomology
- Herpetology
- Ornithology
04 / Record
Published work
249 publications, 1995–2026
Thirty-one years of being the person other laboratories bring their data to. The first entry is a logistic regression for managing zoo neonates; the most recent is multiomic endotyping of trauma patients in Science Translational Medicine. In between: whooping crane studbooks, prairie soil fungi, dilated cardiomyopathy genetics, paediatric glioma, and the antimicrobial resistome of the beef production chain.
Consortium and resource work included — FaceBase, and the MEGARes antimicrobial resistance database.
05 / Software
Open source
The analysis framework behind much of the sequence-based work is released publicly.
Paralome
A reproducible, checkpointed workflow framework for omics analysis on HPC — one backbone under independent RNA, ATAC and variant pipelines.
Read about Paralome →06 / Contact
Ken Jones
Ken.Jones@bioinformaticsolutions.com