Documentation
ClonoLoom opens the output of an annotation pipeline for T- and B-cell receptors, checks it, and groups it into clonotypes by a definition you can see. From there it measures each repertoire with its uncertainty and compares samples, keeping every step on record. Analysis runs on your computer; public queries and sequence handoffs require explicit review.
A path through it
Each page is a subject, in the order an analysis usually goes. Start with Getting started and an example.
| Subject | Pages | What you do there |
|---|---|---|
| Start | Getting started, Opening data | Install, try an example, open your own data or another tool's table, and check them |
| Quality and clonotypes | Quality control, Contamination evidence, Batch QC, Clonotypes | See each sample's quality and the vendor's numbers recomputed, filter rows with a reason, and choose what a clonotype is |
| Repertoires | Diversity, Rarefaction, Genes and junctions, Overlap | Measure each sample's diversity, compare samples at equal coverage, count their genes, junction lengths and amino acids, and measure how much they overlap, each with its uncertainty |
| Tracking | Subjects and time points, Following clonotypes, Expanded and contracted clonotypes, Classes across time points, Attribution to a reference | Give samples their subjects, time points and reference samples, checked before anything is followed; follow clonotypes through each subject's time points, each frequency with its interval; test each clonotype between two time points, with false discoveries controlled, and class it across several; measure each sample against its reference sample |
| Specificity | Motifs and sequence signatures, Subject recurrence and HLA, Similar receptors (TCRdist), Neighborhoods and backgrounds, Generation probabilities, Clonotype networks, Specificity databases | Compare every T-cell clonotype with every other by TCRdist, as tcrdist3 computes it, and find each one's neighbors and nearest, loop by loop; find the neighborhoods a background does not explain; compute how likely recombination alone is to make each clonotype, as OLGA computes it |
| Cells | Expression evidence, Pseudobulk counts, Paired expression | Inspect raw cell evidence and paired gene expression while preserving subject and sample boundaries |
| B cells | Sequence evidence, Families, Germlines, SHM, Lineages, Isotypes | Inspect annotated sequences, choose a family threshold, reconstruct reference projections, measure evaluable substitutions, explore tree uncertainty and summarize observed constant calls; follow the artificial immunization workflow |
| Cohorts | Study design, Subject features, Group comparisons, Replicated counts, Definition robustness | Select biological observations, save designs and features, and compare subjects with explicit tests, effect intervals and provenance |
| Results | Figures and exports, The record | Save figures and export data with their provenance and license; the project, its history and the methods paragraph |
| Reference | Simulated studies, Remote control | Studies whose truth is known, to check methods against; drive ClonoLoom from scripts and notebooks |
How it works
- One program, on your computer. ClonoLoom reads local sources without changing or moving them. Uploads, exports and portable archives create separate copies when requested. Public-repository queries and reviewed sibling handoffs require explicit actions and the documented permissions; opening a view sends no local receptor sequences.
- Checked before it is used. Data are checked before analysis, and the vendors' own numbers are recomputed from the rows. Each method is validated against an independent implementation or a known truth, with the numbers on the Science page.
- The definition travels with the result. Every clonotype count, figure, export and methods paragraph states the clonotype definition and the abundance unit it was made under.
- Everything is kept. Results are kept with their lineage, parameters, seeds and the SHA-256 of their files; asking again answers the result kept. Filters are kept as versions, and undo covers every change.
References
Four references stay in the repository: Formats, every file ClonoLoom reads and writes; the remote control reference, every action a script can send; simulated studies, the model and its truth; and installing. The changelog says what each release added.