about summary refs log tree commit diff
path: root/docs/www
AgeCommit message (Collapse)Author
2 daysoperator checklist: the ravanan build is a patched one, and whyPjotr Prins
Three local patches, the third of which is the reason a 20-chromosome scatter now runs 20-wide instead of 1-wide: {"job": ...} versus {"jobs": [...]}, which is slurmrestd's heterogeneous-job form. With the node declaring 30 CPUs (cold 24.7 s) and the submit fixed (cold 9.4 s), ravanan is within 3x of gemma-wrapper's 3.1 s for the same rows instead of 33x.
3 daysmonitor: run the cached engines cold, not just warmPjotr Prins
A warm run answers from caches the previous binary filled, so it can pass while the cold path -- what a new caller's first request actually does -- is broken. `--cold` sends each cached engine a payload with one sample value nudged by a thousandth, then repeats the original: every engine here keys its cache on content (gemma-wrapper's hash, ravanan's interned inputs, rqtl/qtl2/bulklmm's own), so the first is a miss and the second a hit. Both runs are judged; the line reports "cold X / warm Y". Measured on production with the 30-sample fixture: compute-ravanan 102.0 s cold / 3.7 s warm compute-gemma 3.1 s cold / 2.0 s warm compute-rqtl 7.5 s cold / 7.3 s warm compute-qtl2 3.3 s cold / 3.3 s warm compute-fitqtl 6.7 s cold / 6.9 s warm compute-bulklmm 19.6 s cold / 19.0 s warm ravanan is the one where the cache is the difference between minutes and seconds; the rest are compute-bound at this size, and the honest thing is to print the pair rather than assert a speedup. A check whose payload cannot be perturbed skips with that reason instead of reporting two warm runs as cold and warm. The operator checklist gains the cold run as part of deploying, with both costs named: ravanan's cold run takes minutes, and every cold run adds cache entries.
3 dayscorrelation: the matrix endpoint, and the rule that a deploy ends with a ↵Pjotr Prins
monitor run `POST /compute/correlation/matrix` takes a list of `{dataset, trait_name}` references -- a collection, as GN2's matrix form posts it, which spans datasets -- and answers an N x N matrix of r, p and n. The five decisions that are not obvious, each recorded where a reader will look for them: * **One method for the whole matrix.** GN2's `show_corr_matrix.py` flips to Spearman mid-row once a Pearson r passes 0.999, which the self-pair always does, so its matrix is asymmetric by construction. Here the method is chosen once and r[i][j] == r[j][i] exactly. * **Each unordered pair is computed once and mirrored** -- N(N-1)/2 engine calls, not N^2. * **The diagonal is by construction**: r = 1, p = null, n = the trait's own count, so half the traits cannot come back as 1.0000000000000002 with a NaN p. * **Per-cell n is its own matrix**, and a cell under `min_overlap` (default 6) is null in r and p while its n still says how many points there were. GN2 accepts 2, where |r| = 1 by definition. * **The sample axis is the first trait's group** (#106), resolved once and stated in the reply with the number of individuals it kept. Fetches are one query per dataset (`get_dataset_values_for_traits`, `get_publish_values_for_traits`), not one per trait: a matrix chooses a handful out of twenty thousand. Capped at 1 000 traits, because the cost is quadratic in the correlations and in the reply. No PCA -- the r matrix has to be trusted on its own first. Verified twice: cells agree with scipy to the last bit (BXDPublish 10001/10002/10005: 0.6756270426459695, 0.6950450737755965, 0.8819879006700077 against scipy 0.6756270426459695, 0.6950450737755964, 0.8819879006700073), and a cell equals what the row endpoint says for the same pair (0.6950450737755965 in both). Edge cases exercised: a 1x1 matrix, `min_overlap` above every cell's n, an unknown trait, `bicor`, and no traits at all. The operator checklist now says a deploy ends with `bin/monitor-gn4 --tier all`, and says why: `/version` can report a commit and a reachable database while every engine that shells out is dead. The monitor gained a matrix check whose judge enforces square, symmetric and a unit diagonal.
3 daysmonitor: a --filter for "just retest this", and track the toolPjotr Prins
`--only` needs exact check names, which is no help when what you know is the dependency ("julia", "Rscript", "slurmrestd"). `--filter` matches a substring against the check's name, the sentence describing what it needs, and its hint -- so `--filter julia` finds compute-bulklmm even though its name says nothing about Julia. It selects tiers like `--only` does, and says so when nothing matches *because* discovery has not run (the compute checks do not exist until a group with samples is known). Also un-ignores bin/: the monitor is source, and an ops tool that is not versioned is the problem ai-memory #38 already named. Bytecode stays out. Co-report: the whole compute tier is green now (9 ok, 0 failed), including BulkLMM -- the recipe and the depot quirk are in the operator checklist.
3 daysoperator checklist: the tool block env.sh must keep, and why BulkLMM cannot ↵Pjotr Prins
work yet Writing down what the monitor found and what fixed it, because it is config that lives outside the repository and is therefore easy to lose at the next deploy: with only the cache dirs set, every engine that shells out answers 502 while the whole read surface and search look healthy. Measured on production after the fix (bin/monitor-gn4 --tier compute): 8 ok, 1 failed -- correlation, gemma (21 055 rows), rqtl, qtl2 (20 453), fitqtl, ravanan (21 055 rows), bnw, bnw-examples. Before: 1 ok, 7 failed, 1 busy. The remaining failure is BulkLMM and it is not a config line: the Julia project is registry-based and no depot on this box has its packages, so it stops at "Run Pkg.instantiate()". Populating a depot downloads from the Julia registry, which should be a decision rather than a side effect.
3 daysoperator checklist: how production actually runs now (store binary, no ↵Pjotr Prins
container) The layout, the restart command, where mutable state lives, and the two mistakes that cost time today: the CWL comes from the checkout rather than the store, and only two of the four ravanan 0.2.0 builds on this box convert --slurm-api-endpoint to a URI -- the others die inside slurm-api.scm with struct-vtable on a string, after staging every input, which reads like a Slurm fault rather than an argument one.
4 dayslog: name the person and the credential, and resolve handles from the registryPjotr Prins
The access log said who (`user`) but not with what: with short tokens a request can be made by a shared credential, so "which person" and "which credential" are now separate fields -- `scheme` ("jwt" | "opaque") and, for an opaque token, `token` (its registry handle, e.g. `agent-ravanan-3`). Never the token or its hash, and the path still carries no query string. The line also gains `ts` (RFC 3339 UTC), so a redirected stderr file is self-describing rather than relying on journald's prefix. Hand-rolled civil-from-days rather than a new crate: one format, and no Cargo.lock drift. scripts/gn4-log.py reads such a log and shows it as a table with the handle joined against tokens.sqlite: the subject as the registry has it now (flagging a handle whose subject has been repointed), and the token's status today -- [revoked], [expired], [unknown] -- which is how "an agent was working at 14:02" is told apart from "an agent was working at 14:02 with a credential revoked since". It reads the registry read-only, passes non-JSON lines through in table mode (the process log interleaves backtraces), and offers --json and --summary for jq and for counting. CallerSlot now carries the whole Caller (scheme, id, subject) rather than just the subject string.
4 daystokens: short opaque API tokens in a single tokens.sqlitePjotr Prins
A second credential format beside gn-auth's JWT, for agents and scripts: `gn4ai_` + 56 body characters + 2 checksum characters = 64, from an alphabet with no look-alikes, so a typo is refused before the registry is consulted. Opaque by design -- nothing is stored in the token; every fact lives in one SQLite database (`tokens.sqlite`, GN4_API_TOKENS_DB, mode 0600, gitignored) keyed by the SHA-256 of the secret. Revoking is a row, takes effect on the next request, and a token in a log line reveals nothing; lose the registry and every short token stops working. Administration is local-only: GET/POST /tokens, DELETE /tokens/{id} and GET /tokens/who are behind a router layer that refuses any caller whose TCP peer is not 127.x.x.x/::1, and any request claiming to come through a proxy. A layer rather than a per-handler check, so nothing of a remote request is parsed before it is refused; there is no configuration switch, because minting credentials must never be reachable from the network. A short token reads public data only: every token-only endpoint, and `#[gnauth(secured)]` endpoints for public rows; a confidential dataset is refused with a 403 that names the JWT route, since gn-auth cannot validate an opaque token (the on-the-fly-bridge and registry-grant alternatives were both rejected for v1). scripts/gn4-token.py create|list|revoke|who drives the routes, so there is one implementation of the format and the revocation rules; the token is printed exactly once. Tests: the localhost guard over all four routes (remote peer, no peer, proxied header), and mint -> authenticate -> who -> revoke -> 401.
4 daysskills: ravanan is the default GEMMA runner, gemma-wrapper is the referencePjotr Prins
docs/www/SKILL.md: the workflow example now calls /compute/gemma/ravanan, and a new "Which GEMMA endpoint" section says when not to: Slurm down or submission failing, the user asks, a field ravanan refuses (covariates, normalization), or verifying results -- where the wrapper endpoint is the reference and scripts/gemma-compare.py is the check. docs/dev/SKILL.md: the same rule from the developer side, in Testing -- verify mapping changes by sending one body to both endpoints and never 'fix' a difference by switching endpoints, plus the reminder that the wrapper's cache only hits from 5eb53fa on, so a stale gemma-wrapper on PATH looks like a broken cache.
4 dayslogs: one access line per request, naming the callerPjotr
The proxy logs the caller through x-gn4-user, but the process log -- the one an operator tails while a job runs -- named nobody, so "who started this GEMMA run" meant going to the nginx log or guessing. genenetwork4 now writes one compact JSON line per request on stderr: {"event":"request","method":"GET","path":"/groups/BXD","status":200, "took_ms":10,"request_id":"…","user":"zachs"} Decisions in it: `user` is absent, not "-", when no token was verified (a dash is a value a reader can mistake for a user name); the path is logged without its query string, because a query can carry a credential and has no business in a log; the token value itself is never logged, only its subject; and the line carries the `request_id`, so it sits beside the existing `event: error` lines for 5xx and one grep joins a failure to the request that caused it. `GN4_ACCESS_LOG=false` turns it off where the proxy's log is the record of truth. Verified on a probe instance: with the test token the line says user "gn4-test-stub", with zachs's token user "zachs", and an unauthenticated 401 has no user field at all. docs/www/operator-checklist.md claimed "genenetwork4 does not log the token's sub" -- true until now, which is exactly the kind of sentence that rots; it is replaced by the line above, what it contains and what it deliberately omits. README's x-gn4-user paragraph gains the same, next to the identity it already described.
4 daysdocs: always offer one sentence of what to do next, unless disabledPjotr
The API already computes every reply's options -- `next` are GETs that answer 200, `actions` are the POSTs to run next -- and a caller that reads them learned nothing extra from being told "you could now do X" in prose. What was missing was the instruction: an answer that stops at the result leaves the reader to work out the graph. docs/www/SKILL.md gains "Always offer the next step": end every answer with ONE sentence built from that reply's own next/actions -- after a compute run, a lookup, a plot, an upload, a basket edit. Prefer the action href over a hand-built call. A refusal is an answer too: its next or detail names the alternative (bulklmm for a GEMMA permutation, /compute/gemma for a covariate fit). Do not stack suggestions; one next plus at most one genuinely different alternative. "Skip the suggestions" / "just the result" turns it off for the session -- it is the sentence that is disabled, not the content. And never propose a step the API cannot do: if nothing in next/actions fits, say so. README states the same bargain where the next-link promise is described: the server's half is that the suggestion is always reachable, the assistant's half is that it is one sentence and it is a link the reply actually carries.
4 daysindividuals: select who enters a mapping, by namePjotr
Parents and F1s are excluded by default (as in GeneNetwork2/3), with include_parents_f1 to keep them all. That boolean was not enough for what was asked -- "include these, drop that substrain" -- and the exclusion list itself is data (parents_and_f1s.json, seven groups) that no request can reach. F1 names also differ per group (B6D2F1 for BXD, BHF1/HBF1 for BXH, C57BL/6JxBXD037F1 for the intercross F1s), so pattern matching is the wrong tool. Every group-mapping endpoint -- /compute/gemma, /compute/gemma/ravanan, /compute/rqtl, /compute/qtl2, /compute/fitqtl, /compute/bulklmm -- now takes "individuals": {"include": [...], "exclude": [...]} applied after the parent/F1 rule, exclude winning over include. An unknown name is a 422 invalid_individuals whose message lists the panel: a silently dropped animal changes a kinship matrix without anyone noticing, so the failure has to be loud. genotypes::select_panel is the one implementation, and it returns what it did (dropped parents/F1s, re-included by name, excluded by name) so the Genotype note can say it: "197 individuals, 38 parents/F1s dropped; 1 excluded by name". The note now also spells out the distinction that makes this option treacherous: the kinship uses *every* individual in the genotype file, the association uses only those with a trait value -- GEMMA reads no phenotypes for the kinship. So "include the F1s" changes the relatedness structure without putting the F1s in the fit, and a request can legitimately have 234 individuals in the genotype file and 34 in the association. Verified on BXD.8: excluding BXD1 gives 197 individuals / 33 values; include_parents_f1 with an F1 re-included and a substrain excluded gives 234 / 34. A basket supplies values by name, so a selection applies to a basket-fed mapping the same way: the individuals excluded are simply not in the panel any more, and their basket values are ignored. GenotypeError gains Invalid for this (422) so a caller's typo is not reported as a missing genotype file; the read-only endpoints that can never see one map it the same way for exhaustiveness.
5 daysbnw: module scores as trait values, and BNW's worked examplesPjotr
Two additions that make the endpoint usable rather than merely correct. POST /compute/bnw/features -- from a network to a mappable trait. Every node with at least one parent at the threshold is a module (the node plus its parents), scored per sample by the first principal component of the module's standardised members, sign-corrected so the score rises with the node. Without ?module= the reply is one row per module -- module, members, size, variance_explained -- where variance_explained says whether the score is a fair summary at all. With ?module=NAME it is that module's per-sample scores in the {name, value} shape every mapping endpoint already takes, so the chain runs network -> module -> /compute/rqtl or /compute/gemma with no translation layer. That is what BNW is for in GeneNetwork; the diagram is a by-product. The notes name the construction (a citable module score must say what it is) and repeat the caveat: a module is a correlation structure, not a QTL. GET /compute/bnw/examples and /compute/bnw/examples/{name} -- BNW's own worked examples, for an agent to explore instead of inventing data. The five datasets its tutorial and examples pages present are listed with their sizes, and each is served as a matrix in the values shape /compute/bnw takes, with a ready-to-post action carrying the body. ?columns= restricts the matrix (the spleen example has 14 variables and a caller may want fewer). genotype_column warns that a column is a genotype code: this driver is continuous-only, so it enters the network as a number, the way the web application's GeneNetwork path treats it. Files come from BNW_EXAMPLES_DIR (a gn-bnw checkout's sourcecodes/ example_datasets); a deployment without it answers 404 no_examples. Measured on the dev instance: features over five BXD traits gives two modules (BXD::10001 with BXD::10003 and BXD::10041, 60.4% of variance; BXD::10044 with BXD::10001, 27.2%), and BXD::10001's 19 scores posted straight to /compute/rqtl and /compute/gemma both return 200 -- gemma peaks on chromosome 6. tests/bnw.rs covers the module list, one module's scores in mapping shape, the unknown-module 422, the example list, one example's matrix and its ready-to-post action, and a 404 for a name that does not exist.
5 daysplot/bnw: two SVG figures, drawn the way /plot/gemma draws onePjotr
Post the reply from /compute/bnw back and get an SVG: the same plumbing as /plot/gemma and /plot/bulklmm (plot::render, PLOT_PYTHON + BNW_PLOTTER, image/ svg+xml, the plotter's line in x-plot-summary), nothing recomputed. Two figures, because one cannot say what the other can: * the network (default): nodes on a circle, one arrow per edge that survived the threshold, arrow width and label = the edge's posterior, and the below-threshold edges as faint dotted lines -- so a reader sees what was *considered* as well as what survived, which is half of why the posterior table exists; * ?kind=posterior: every candidate edge as a cell in a matrix, dark = high posterior, outlined cells are the ones above the threshold. The diagram cannot show that an edge came within 0.02 of the threshold; this can. Pure-Python SVG, standard library only: no graphviz in the deployment. A circular layout is less pretty than a force-directed one and is deterministic, which matters more for a figure that accompanies a JSON result. Node labels keep the distinguishing tail of a long id, with the full name beside the node. Verified live on the dev instance: POST /plot/bnw with a saved /compute/bnw reply returns 200, image/svg+xml, 5 256 bytes, and x-plot-summary naming the strongest edge (Trait2 -> Trait1, posterior 1.000); ?kind=posterior returns the second figure. tests/plot.rs covers both figures plus the 422s (not a BNW result, unknown kind).
6 dayscompute/bnw: a Bayesian network over traits, called out to BNW's driverPjotr
POST /compute/bnw learns a network over a set of variables and returns one row per candidate edge: from, to, posterior, in_structure. Inputs are published traits of a group (group + traits) or an explicit matrix (values: one entry per variable with per-sample values), which covers the case the database cannot -- a basket's columns, a handful of probe sets. BNW's knowledge stays in the gn-bnw repository. GN4 writes a variables x samples matrix, runs `bnw learn` (BNW_CMD, BNW_BIN_DIR, BNW_CACHE_DIR) and parses the reply: nothing about the three C programs, the private work directory or the type line in the input file lives here, so BNW's pipeline can change without a GN4 release. The driver's failures map onto the API's own: bad input 422, a missing driver 502 bnw_not_installed (its exit 77), a failed computation 502 bnw_failed in the driver's words. Decisions worth recording: * parameters are BNW's own defaults -- maxparents 4, k 1, tiers 3, threshold 0.5, read from its create_tiers_gom_part1.php rather than chosen here -- and the reply names the values used, because a network at 0.5 and one at 0.9 are different answers to the same question; * at most 40 variables and at least 3 samples complete across every variable. The first is the structure search's exponential in-degree; the second is BNW modelling a complete multivariate normal, so a sample missing any value is dropped -- and the note says how many survived rather than silently shrinking the panel; * its own gate, always claimed (threshold 0): there is no such thing as a small BNW run, the cost is the search, not the number of individuals; * the notes say plainly that a posterior is not a p-value and that a network is not a QTL, because that is the easiest thing to over-read. tests/bnw.rs compares the endpoint against the answer the BNW *web application* gives for its own six-trait example (the same fixture as the driver's parity check, gn-bnw/test/golden/): the five edges must match, and the notes must name the parameters and carry the caveat. It skips when the driver or its binaries are absent. Not deployed to production.
6 daysdocs: say that project memory is how a session recovers, and where it livesPjotr
The ai-memory store already existed; what was missing was any mention that a reader should use it before starting work, and what they would find. Two concrete things a cold reader needs and cannot reconstruct from a diff: an operational RUNBOOK entry (the Slurm container from gn-machines, its fixed JWT, the environment /compute/gemma/ravanan needs, the four prerequisites that bit the first time, and the fastest test: one chromosome, 14 s), and the "why is this like this" entries behind endpoints that exist because something else was wrong (-pve reporting, the next-link graph, the reference-genome linkage). docs/www/SKILL.md gains section 7a with the three query forms, the topic names that exist (ravanan-slurm, reference-genomes, gemma-pve) and both measured costs; README gains the same note under Development, since that is where a maintainer looks first. States the limit honestly: the store is local, gitignored and per repository -- it survives a cleared session, not a lost machine, which is why every conclusion is also in a commit message and in these docs.
7 daysgemma: the pve note carries every chromosome, and the log carries the numbersPjotr
The note reported the range and named the two chromosomes that bounded it, which is not enough to answer the question the numbers exist for. Under --loco each pve belongs to a *different* kinship matrix, and GEMMA's per-chromosome .log.txt files live on the server -- so a caller told "0.31 to 0.95" could not ask which chromosome was which, or compare a single chromosome against BulkLMM. The note now lists all twenty, compactly, with the range and the two extremes keeping their vg/ve: GEMMA -pve, null model, per chromosome (LOCO): chr1 0.6674±0.3451, chr2 0.8426±0.3668, ..., chrX 0.8563±0.3525. That is pve 0.3118 to 0.9549 over 20 chromosomes; lowest chr8 pve 0.3118 ± 0.5009 (vg 23.3659, ve 12.3389), highest chr7 pve 0.9549 ± 0.3570 (vg 76.0933, ve 0.8661). and streaming::append_lines puts the unrounded values in the run log as variance: chr7 pve=0.954938 se_pve=0.357032 vg=76.0933 ve=0.86612 so a ?id= run's /compute/gemma/status stream carries them while the run is still in flight, rather than only in a reply someone has to keep. Both measured on BXD::10001 through the dev server before committing. The values are printed with four significant figures (scientific when the magnitude needs it), unchanged: rounding variance components to four decimals was the earlier bug, and these are the same numbers a reader compares with BulkLMM's h2. Docs: README's /compute/gemma section, SKILL's gemma bullet ("quote per-chromosome values, not an average: they are twenty different kinship matrices") and docs/bulklmm.md's example.
7 daysdocs: how pve gets deployed, and what the deployed box runsPjotr
GEMMA_WRAPPER_CMD is the lever: the profile's gemma-wrapper may be older than the pin, and pointing the variable at the store path is enough -- -pve reaches GEMMA either way, but the per-run .log.txt files where pve lives are only kept beside the archive from 5eb53fa on. Recorded the verified production state too: both processes restarted with gemma-wrapper-1.00-pre1-4.5eb53fa, and POST /compute/gemma for BXD::10001 on the running instance returns pve 0.3125 to 0.9558 over 20 chromosomes -- measured there, not on a fixture.
7 daysgemma: report GEMMA's own pve for the null modelPjotr
The GWA step now passes -pve, and the reply gains a note with the variance components GEMMA computed for the null model. Under --loco there is one per chromosome, so the note gives their range and names the two that bound it: GEMMA -pve, null model, per chromosome (LOCO): pve 0.3118 to 0.9549 over 20 chromosomes -- lowest 8 pve 0.3118 +/- 0.5009 (vg 23.3659, ve 12.3389), highest 7 pve 0.9549 +/- 0.3570 (vg 76.0933, ve 0.8661). That is BXD::10001 measured, not assumed. pve is the proportion of variance the kinship explains -- the same quantity as /compute/bulklmm's h2, per chromosome rather than genome-wide -- which is what the 16627/16769 comparison against BulkLMM needed. Reading the numbers took three decisions worth stating: - Four decimals is wrong for variance components. The first live run reported 'vg 0.0000, ve 0.0000' for a BXD::16769 null fit whose real components are 3.411e-5 and 3.411e-10 -- both rounded to zero, which reads as 'no variance' when the split is 1.0. Components are printed with four significant figures, scientific when the magnitude needs it, the way GEMMA's own logs do with %g. - se(pve) is -nan on a boundary, and Rust parses -nan as a NaN f64, so the numeric parse alone let it through as NaN. Non-finite values are now 'not available': a boundary estimate has no standard error. - Twenty chromosomes of values bury the other notes, and every value is in the .log.txt beside the archive anyway, so the note names the ends. Requires gemma-wrapper 5eb53fa, which keeps each run's .log.txt beside its archive instead of only inside the .tar.xz. Docs: README's /compute/gemma section, SKILL's gemma bullet (compare pve with h2, never vg with sigma2_e -- pve is scale-free), and docs/bulklmm.md's 'what would settle it', which now says the plumbing is in rather than pending.
7 daysgenes: the alias reply carries the entity's label and its gene link-outsPjotr
The GTEx problem -- it is human-only, and a mouse symbol there is an empty report -- is answered by the endpoint that already knows which entity it matched: /genes/{symbol}/aliases now returns label the entity's canonical symbol in that species ("Shh" for mouse, "SHH" for ?species=human) -- the symbol a human-only service needs links [{name, href, description}] built for that symbol and species: RGD, MGI, BioAlliance, NCBI Gene, Wikidata (the QID), UniProt, STRING, GeneMANIA and GeneCup, plus GTEx/Open Targets/Protein Atlas/GWAS Catalog for a human entity or BioGPS/Allen Brain Atlas for a mouse one So a mouse Shh reaches GTEx through the human ortholog's label rather than a guessed one, and a client no longer holds the URL vocabulary. Only links that answer are built: HomoloGene, UniGene, Genotation, GeneBridge and PhenoGen are gone, and Panther is browser-only (see SKILL.md "Gene link-outs", where each URL was checked). A cached entity written before the label was kept is refreshed rather than served -- otherwise the link-outs would be built on a symbol it does not have -- and `stale: true` still says when Wikidata was unreachable and the cached answer was used. tests/wikidata.rs: a mouse gene gets the mouse services and no GTEx; the human ortholog gets label SHH and a GTEx href built from it. Both halves skip when the reply is stale, since a Wikidata timeout is not a failure.
7 daysdocs: the gene link-outs GN2's trait page offers, checked and tabletedPjotr
GeneNetwork2's webqtlConfig.py holds the link-out URLs its trait page shows by gene symbol; twelve of them were still alive when checked (2026-11-11, HTTP 200 and, where the page is not a JavaScript shell, the gene found), and the SKILL now carries them as a table with descriptions and the URL templates to hand a user: RGD (gene search, mouse/rat/human -- the best single hub for both rodents), MGI, Alliance of Genome Resources, NCBI Gene, Wikidata (the QID /genes/{symbol}/aliases already returns), UniProt (symbol search or accession), Ensembl (gene id), GTEx (human symbols only), STRING, GeneMANIA (per species), Open Targets, Human Protein Atlas, GWBI GWAS Catalog, BioGPS, Allen Brain Atlas, GeneCup. Named so they are not offered by mistake: HomoloGene (NCBI retired it 2021), UniGene (410), Genotation and GeneBridge (dead), PhenoGen (404), chibi GEMMA, wiki-Pi; PANTHER answers 403 to a bot but works in a browser. Two conventions: quote the species the symbol came from, and pick the species-appropriate service (a mouse symbol in GTEx is an empty report). The API hook is noted but not built: a `links` array on /genes/{symbol}/aliases, so a client need not hold the URL vocabulary.
7 daysdocs: say that /plot/gemma draws no line unless one is asked forPjotr
The OpenAPI description still promised "a Bonferroni reference line by default", and GEMMA's own suggested plot call said the same. Both now say what the code does, and where the number comes from: the plotter computes 0.05 over the analysed markers -- conservative under linkage and unrelated to relatedness -- and ?thresholds= takes one computed elsewhere, each line labelled with its source. Also corrected in the README's /plot section and in SKILL.md, where the same promise had been copied.
7 daysgemma: the permutation refusal names both routes, including running GEMMA itselfPjotr
We do not run GEMMA permutations -- each repeats the whole LOCO scan and the gate gives GEMMA one slot -- but "not supported here" is not the same as "not available". The refusal now says both: 1. POST /compute/bulklmm with nperm: 1000 -- an LMM whose thresholds come from rotations that keep the kinship (R/qtl's nperm does not, and is anti-conservative on a structured panel); 2. run GEMMA's own permutation locally. The reply's notes already name the bimbam pair and the phenotype file, so `gemma-wrapper --permutate` or `gemma -lmm 9 --permutate` on those inputs reproduces exactly the null GEMMA would have used -- hours instead of seconds, which is the honest price of that particular null. docs/bulklmm.md gains the table this belongs in: three ways to get a threshold, ordered by how honest the null is for a structured panel (rotation with kinship; GEMMA's own LOCO permutation run locally; R/qtl's nperm as a sanity check only). SKILL.md says the same in two lines. tests/gemma.rs now asserts both halves of the refusal, and the earlier assertion on /compute/rqtl -- which stopped being true when the refusal was rewritten -- is replaced rather than left passing by accident.
7 daysbulklmm: exact marker-block parallelism, and what the cores are actually doingPjotr
Reported: a BulkLMM run does not look multicore. It is not a misconfiguration. BulkLMM's single-trait `scan()` -- what this endpoint calls -- has no `Threads.@threads` at all; the package threads `bulkscan()` (many traits) and bulkscan_helpers, not scan. The parallelism that was visible before was BLAS (OpenBLAS defaulting to all 96 cores here), which is why `user` time ran 4x `real` while Julia threads made no difference: --threads 1 13.7 s, --threads 4 12.3 s. * The driver splits the markers into independent blocks and scans them on Julia threads (--blocks, default = BULKLMM_THREADS). It is exact: the LOD for each marker is independent given the null variance components, which depend on y and K only. Verified identical -- peak LOD 9.589938 and the same permutation thresholds at 1 and 4 blocks -- and permutation thresholds combine by taking each permutation's maximum across blocks before the quantiles, which is what a genome-wide threshold is. * The parsed matrix and the kinship are cached per panel (read 1.3 s -> 0.4 s, kinship 0.9 s -> 0), because both depend only on the cross. * The summary in `notes` now says what ran: julia threads, blocks, BLAS threads, and the read/scan split. * The hand parser is back: CSV.jl into a DataFrame is an order of magnitude slower for a 21056 x 235 matrix (14-17 s against 2 s). * BLAS is capped at 4 threads -- 96 was slower (10.5 s against 7.7 s for one scan). Measured on BXD.8 trait 16627 (21056 markers): 1 block 20.9 s, 4 blocks 18.3 s, same answer. The gain is modest because the LMM loop is ~8 s of the run and each block re-estimates its variance components. Where the cores pay is bulkscan() -- many traits in one process, threaded over trait blocks -- which is the shape the next endpoint wants; documented in the README rather than left as folklore.
7 daysbulklmm: an LMM scan that can be asked for thresholdsPjotr
GEMMA's LOCO scan is the reference for a biallelic panel, but its p-values have no threshold we can compute: a GEMMA permutation repeats the whole scan, which is why /compute/gemma refuses nperm. R/qtl's nperm is the usual stand-in and it is the wrong null -- it shuffles phenotypes as if the individuals were unrelated. BulkLMM.jl fits an LMM per marker with a global kinship matrix and permutes by rotating the data through the kinship's eigen decomposition, so a threshold costs seconds. POST /compute/bulklmm {group, genotype_version, samples|basket, nperm, seed, reml, assumption} Rows are marker, chr, pos (Mb), lod, beta, h2. BXD trait 16627 (BXD.8, 35 phenotyped animals, 21056 markers) in 22 s: peak rsm10000000888 chr2 4.92 Mb, LOD 9.59, and thresholds alpha 0.10 -> 3.51, 0.05 -> 3.81, 0.01 -> 4.47 from 200 rotations with seed 42. That is the trait's published locus and LOD, and a significant peak against a threshold this API could not compute before. Two differences from GEMMA are in every reply, not implied: a global kinship matrix rather than leave-one-chromosome-out, and h2_null saying how much relatedness was found (here ~0, so the fit is effectively fixed-effects -- which is why it agrees with R/qtl where GEMMA differs). * src/bulklmm.rs prepares what Julia reads -- the BIMBAM dosages as a strains x markers matrix (our BIMBAM pair is already the allele dosages BulkLMM wants), a marker map, the phenotype and covariates -- runs the driver through streaming::run_streaming_with_env, and reads the CSV back. A row whose LOD is NaN is dropped: NaN is a valid f64 in Rust and would otherwise be reported as a value. * scripts/bulklmm_scan.jl is the driver (scan/calcKinship/get_thresholds), with the seed recorded so a threshold set is reproducible. * gate::bulklmm (2 slots, GEMMA-sized threshold): a permutation here is cheap, unlike R/qtl's. * Julia comes from the deployment's own store binary with a prepared depot -- 79 packages, 139 MB, no GPU/CUDA stack -- and every request runs JULIA_PKG_OFFLINE=true, so a scan can never download anything. `--project=<path>`, not `--project <path>`: Julia 1.8 takes the space-separated form as a program argument. tests/bulklmm.rs: the peak stays on chr2 above LOD 3, the thresholds and seed are in the notes, and a scan without permutations says there is no threshold rather than leaving a bare LOD to be read as significant.
8 daysqtl2: name the genotype model, and say what the conversion droppedPjotr
Two engines reading one riset panel can fit different predictors, and the reply now says which one it used and what it cost -- with this panel's own numbers. * `genotype_model: "riself" | "f2"` on /compute/qtl2 (`het` kept as the older spelling: drop/keep). riself is the default: R/qtl's convert2riself coding, the H calls become missing and the HMM re-imputes them. f2 keeps them as a third genotype -- a diagnostic. * The conversion now counts the calls it drops and keeps the count beside the cross (conversion.json), so a reply can say "610684 heterozygous call(s) (12.7 %) were dropped". A cross converted before that file existed measures the count once, on first use, rather than never. * Crossed with the note that already explained the H drop, so the two read together: what riself does, how many calls it cost, and what f2 would give. Documented in the README as "Why R/qtl and R/qtl2 disagree on a riset panel", with the measurements (BXD trait 16627, BXD.8: R/qtl 9.59 at rs33299937; qtl2 riself 3.73; qtl2 f2 9.68), R/qtl's adaptive convert2riself rule quoted from its own source, and qtl2's side: riself has two states, f2 three, and `genril2` -- from qtl2 0.18, "the progeny are inbred lines", cross_info first column the number of generations -- is the honest model for residual heterozygosity and is available in the 0.42 we run. What is missing for it is recorded precisely: a founder-based cross expects individual genotypes coded as founder alleles rather than 1/2/3 (a prototype with numeric codes read every genotype as missing), so founder_geno from the two parental alleles and an allele-coded het remain to be done, with DOL as the counter-example (its H is a collapsed DO, not residual heterozygosity). tests/qtl2.rs: the model is reported (Cross: riself / Cross: f2), the dropped count is in the reply, keeping the calls finds more, and an unknown model is refused rather than ignored.
8 daystraits: genotypes at a confirmed peak, and the plot of themPjotr
Returning genotypes is the one thing this API otherwise never does, so it is fenced and the fence is the feature: GET /groups/{group}/traits/{trait_id}/peak-genotypes?marker=<marker> * the server re-runs the R/qtl scan for the trait -- the same machinery and gate as /compute/rqtl -- and takes the peak from that run, rather than trusting a client that names a marker; * the marker must lie within ?window= Mb of that peak (default 10), or the answer is 422 not_the_peak quoting the peak's marker, chromosome, Mb and LOD; * the peak must reach ?min_lod= (default 3), or 422 not_a_hit; * only animals that have a value for the trait are listed, and only this marker's calls. Nothing else of the matrix leaves. On DOL 10003 that is UNCHS026345 (chr9 75.76 Mb, LOD 5.86) with 192 animals across A 110 / B 8 / H 74 -- the two a class of eight being exactly the thing worth seeing. POST /plot/genotypes draws R/qtl's Fig 4.1 from it: post the reply's `animals` (or its `genotypes` map) with the trait's values and get an SVG with one point per animal, genotype class across, value up, class means marked. Nothing is recomputed -- the plot is of the numbers the client sends, like the other plots. New script scripts/plot-genotypes.py (stdlib-only SVG, like its siblings) and the GENOTYPE_PLOTTER key. tests/peakgenotypes.rs: the peak's calls for animals with values (with the counts agreeing with the rows), the SVG, and a marker on another chromosome refused with the peak quoted. Also keeps /compute/rqtl's error order: naming both a group and a file is "ambiguous" even when the request also forgot its values.
8 daysuploads: every engine maps an upload by name; same name replacesPjotr
Two things asked for, and one bug each. * POST /uploads/genotypes with an existing name now REPLACES it. A genotype is the user's file to keep, not a record the deployment curates; the digest in every reply says which version ran. * genotype_upload on all four engines -- /compute/{rqtl,gemma,qtl2, fitqtl} -- in place of genotype_file/group. The server writes the form each engine reads: a GN .geno for R/qtl and R/qtl2, a BIMBAM pair (new: genoupload::to_bimbam, coding from the base's @mat/@het/@pat, NA for missing, snps as marker/bp/chr) for GEMMA. No user path travels with a request. Verified against the full 3811-marker BXD.4 matrix uploaded as an update: rqtl 3811 rows, gemma 3809, qtl2 3730, fitqtl by marker. The bug worth naming: values given by name were being aligned against a file, and for GEMMA that file is BIMBAM, which has no individual names -- so the phenotype file came out empty and the wrapper failed with a Ruby backtrace that said nothing. GEMMA's values are now ordered by the upload's own individual list before the pair is written. The related fixes: rqtl and fitqtl align named values onto the file's individuals (they previously demanded its column order), and genotypes::align_to_file is the one place that does it. tests/genoupload.rs: mapping an upload on rqtl/qtl2/fitqtl with values by name, and replacing a name (one row in the listing, a new digest). GEMMA's LOCO kinship needs more markers per chromosome than a small fixture has, so it is exercised against a full panel rather than in the fixture.
8 daysuploads: map an uploaded genotype by name, and clean up after itPjotr
Three follow-ups from putting an upload through the whole loop. * /compute/qtl2's `genotype_file` branch only accepted positional `trait_values`, i.e. it demanded the file's column order -- which a caller with an uploaded population does not have. It now takes `samples` (name and value) or a `basket` and aligns them onto the cross by name, leaving individuals without a value as missing. Positional `trait_values` still work for a file whose order is known. * DELETE /uploads/genotypes/{name} now removes the .geno it wrote on demand: a file nobody can name is a file that confuses the next operator. * scripts/example-genotype-upload.py spreads its markers across chromosomes. One chromosome of markers is a poor example -- a leave-one-chromosome-out kinship has nothing to leave out, so the scan came back with no rows at all, which looked like a broken upload and was really a property of the example. 20 markers over the genome map cleanly (19 rows, the 20th used for the HMM), and the README says to use "kinship": "overall" if a file really has one chromosome.
8 daysuploads: a genotype matrix that updates an existing populationPjotr
Tab-delimited, header names the individuals from column 4, then marker, chr, pos (Mb). Named $USER-GENONAME, stored as text in the owner's own SQLite file -- nothing goes near MariaDB -- and written out on demand as a GN .geno carrying the base's @-metadata, so the R scripts read exactly what was uploaded. An upload updates a population, it does not invent one, and the checks say so: * the base version must exist (resolved through the same genotypes::resolve_named the rest of the API uses); * at most 50 individuals the base does not have, and at least one that it does -- no overlap is a different population, not an update; * at most 200 000 markers, and the refusal says the sysadmin has to organize a file that size, because this is a request path, not an ingest; * the values must be the same kind as the base's: letters, whole numbers or probabilities. A letters panel holding 0/1 would be read as the wrong thing silently, so it is refused naming both kinds. GET /uploads/genotypes lists them, /{name} its metadata (overlap, new individuals, kind, digest), /{name}/markers pages the marker table, and /{name}/geno writes the file and returns the path for `genotype_file` on the compute endpoints. DELETE removes one. scripts/example-genotype-upload.py turns any .geno into a valid update -- a subset of markers, a few individuals dropped and added -- and tests/genoupload.rs builds its own from the deployment's BXD.4 for the same reason: upload, page the markers, write the .geno and check the layout, delete, then break each rule (no overlap, numbers in a letters panel, commas) and check the refusal says which. Unit tests cover the three value kinds, missing cells, ragged rows, a non-numeric position, the 50-new-individual limit, and the .geno layout (Chr first, then the marker -- GN's order, not the upload's).
8 daysmatrices: upload new traits and covariates as one named tablePjotr
The other half of changing data: bringing new columns in. A matrix is a delimited table -- a header row, then one row per individual, the first column naming them -- stored in the uploader's own SQLite file and checked against an existing genotype version's panel before it is accepted. * POST /matrices {name, genotype_version, contents, kinds?}: parse the table (tab, comma or semicolon, guessed from the header), then check it against that version's individuals. The count must match and every name must be in the panel; 92 of 93, or a name the panel does not have, is a 422 that says which. A mapping that silently drops individuals is worse than an error. * kinds marks columns that are covariates; everything else is a trait. * Each column becomes a basket named <matrix>.<column> -- a dot, not a slash, because a basket name has to survive a URL path segment -- so a column maps on any engine with {"basket": "my-traits.bw"} and inherits the revision history and digest without a line of new mapping code. * GET /matrices lists them, GET /matrices/{name} their columns and basket names, GET /matrices/{name}/values the table back (parsed values, null where missing), DELETE removes the matrix and its columns. Two bugs this found in the basket store, both real: * SQLite has foreign keys OFF by default, so the ON DELETE CASCADE never fired: deleting a basket left its rows behind, rowids were reused, and the next create hit "UNIQUE constraint failed: basket_version.basket_id, revision". The pragma is now set, and deletion removes the dependants explicitly, because files written before it have orphans. * The first attempt named a column's basket "<matrix>/<column>", which the basket-name rule (rightly) refuses. The dot keeps both rules. scripts/example-matrix.py writes a table from a genotype file's own header with deterministic values, so the fixture passes the panel check by construction -- tests/matrices.rs builds its own for the same reason and walks the loop: upload, read back, map a column with R/qtl2, delete, plus the three refusals (too few individuals, a name not in the panel, an unknown version).
8 daysbaskets: every revision kept, files named for the user, all four enginesPjotr
Three things the first cut was missing, and both were asked for. * An automatic version history. Revision 1 is the column as first stored and each PATCH adds one -- with its own digest and Unix timestamp -- and earlier revisions are kept. GET /baskets/{name}/revisions is the history, ?revision=N reads an old column, and a mapping can pin one: {"basket": "wbc-edit@2"}. That is the point of a digest: a result quoted yesterday can still be reproduced after the column has moved on. * Files named for the user, not a hash of the subject: BASKET_DIR/<user>.sqlite (0600, WAL). Tokens rotate, subjects do not, so a new token finds the same baskets -- and an operator can see whose baskets a file holds. A subject that is not filename-safe is sanitised and disambiguated with a short digest, and a pre-naming hash file is renamed on first use rather than orphaned. * /compute/rqtl and /compute/gemma and /compute/fitqtl take `basket` too, so all four engines share one meaning: a basket resolves to the same (name, value) pairs a client would have posted, and every reply records the revision and digest it ran. The schema upgrade is a real one, not a reset: an older file's single column becomes revision 1 of the history, cells copied across with the digest and timestamp they already had. Tested by building the old schema by hand and opening it (an_older_file_is_upgraded_with_its_column_intact), which is also how I found that the migration was recreating a table the base schema had already made. Verified on all four engines, head and pinned: rqtl 3811 rows, gemma 3809, qtl2 3730, fitqtl 3 markers -- each reply carrying "Basket '...' revision N (digest ...)".
8 daysbaskets: a user's own column of values, in their own SQLite filePjotr
The next phase is changing data, and the constraint is that MariaDB is never written to. A basket is the storage half of that: take a trait's values, edit some of them (or build a covariate column that does not exist), keep the column under a name, and map that instead. * POST /baskets stores a complete column; GET /baskets lists yours; GET /baskets/{name} its metadata; GET /baskets/{name}/values the cells; PATCH /baskets/{name} replaces the column with a new revision; DELETE removes it. * POST /compute/qtl2 takes "basket": "<name>" in place of `samples`, and both at once is a 422 rather than a silent choice. The mapping code below it is untouched -- a basket resolves to the same (name, value) pairs a client would have posted. * One SQLite file per user: BASKET_DIR/<sha256(token sub)>.sqlite, 0600, WAL, busy_timeout. `GET /baskets` only ever lists the caller's own; the file name is a hash, so a directory listing does not name users. * The full column is stored, not a diff -- the caller has the original beside it, and a mapping needs one complete column. * Text with a parsed copy beside it: raw (what was typed), value (the number or NULL), status (ok/missing/bad). "NA" and 0 stay different claims, and a column with something that is neither a number nor a missing marker (NA, x, -, ., ?) is refused with the samples named -- silently dropping it would change a trait without saying so. * A revision and a digest: an edit bumps the revision and changes the digest, and a mapping reply records both ("Basket 'wbc-edit' revision 2 (digest 9559a70e...)"), so two runs on one name cannot look alike. A result cache must key on the digest, not the name. The loop, end to end on BXD trait 10001: 34 values posted as a basket, one bumped by 5 and one marked NA, revision 2 mapped with /compute/qtl2 -- 3730 rows, peak chr8 112.9 Mb, and the reply's note naming the revision and digest. tests/baskets.rs walks that loop (create, list, values, patch, map by name, delete) and the refusals: a non-number named in the error, a name that cannot become a path, an unknown basket, and basket-plus-samples. Unit tests cover parsing (NA/x/-/. vs NaN/inf vs "twelve"), the digest changing on edit, per-user isolation, and name validation. rqtl/gemma take the same 10-line change next -- their `samples` shapes differ, so they are a separate commit rather than a rushed one.
8 daysapi: scans point at the class counts behind their own peakPjotr
The gotcha notes in SKILL.md tell an agent to count the animals per genotype class before believing a peak; a note in a document is only found by whoever reads it, so the replies now say it too, at the peak they just computed. * suggest::class_hint builds /groups/{g}/genotypes/{v}/markers/{peak}/classes?trait={trait_id} -- a real GET with one placeholder, because a scan knows the trait's values, not its PublishXRef.Id. * /compute/{rqtl,gemma,qtl2,fitqtl} add it to `next` after the PheWAS link, naming the peak each reply found and the version the request named (a {version} placeholder when it named none). * The genotype subtree carries the template too: a version's detail, and a concrete marker's own links. * /compute/qtl2 gains the sentence that explains the divergence: on a riself (riset) panel heterozygous calls are dropped before the scan (R/qtl's convert2riself) and the HMM re-imputes them, while GEMMA keeps them at dosage 0.5, so a mostly-H marker cannot peak there and two engines can name different markers of one region without either being wrong. tests/links.rs: a scan's reply offers the class counts at its own peak -- the link names the version and the peak marker the reply reports, and leaves {trait_id} for the caller.
8 daysdocs: four analysis gotchas, with DOL trait 10003 as the worked examplePjotr
SKILL.md told an agent how to call the endpoints, not how to read them when two engines disagree -- which is the question the last round of analysis actually asked. A note in "Gotchas that cost time": * Count the animals in the peak's genotype class before believing it (.../markers/{marker}/classes?trait=<id>). DOL 10003's GEMMA peak is A n=2 mean 1.8 against B n=171 mean 3.95, with its next two hits the same pattern mirrored: a haplotype carried by a handful of extreme animals, not a population effect. * H is dropped for R/qtl2 on a riself panel (kept by GEMMA as dosage 0.5); its HMM re-imputes, so the observation is lost, not the animal. A mostly-H marker becomes nearly monomorphic for R/qtl2 -- 2 A vs 171 B -- which is why the two engines can name different markers of one region. Neither is wrong. * GEMMA's af is mean(dosage)/2 under B=1, H=0.5, A=0, not an allele frequency (0.469 is half of 0.935); allele frequencies come from .../markers/{marker}/frequencies, and allele1/allele0 are the BIMBAM X, Y placeholders, so the sign of beta is not a strain. * Do not compare scores across engines as one scale: GEMMA -log10 p, R/qtl/rqtl2 LOD, qtlreaper LRS (lod = lrs / 4.604).
8 daysgenotypes: a trait summarised per genotype class at one markerPjotr
The question that follows "why do GEMMA and R/qtl2 peak at different markers": what does the trait look like within each genotype class at one marker. The frequencies endpoints report the distribution of calls, not the values behind them, and the answer must not be "serve the matrix" -- this API never does. GET /groups/{group}/genotypes/{version}/markers/{marker}/classes?trait=<id> -> {"A": {"n", "mean", "sd", "min", "max"}, "H": {...}, "B": {...}} The calls are joined to the trait's per-strain values inside the server (the same lookup, authorization and 404s as /groups/{g}/traits/{id}/values) and reduced to those five numbers per class. Nothing in the payload names an animal. A class with fewer than ten animals is called out in notes, because a peak resting on two is a different claim from one resting on a hundred. It settles the DOL trait 10003 question. At GEMMA's top marker JAX00418132 (chr16 36.35 Mb) the classes are A n=2 mean 1.8, H n=20 mean 2.53, B n=171 mean 3.95 -- the A class is two animals and they are the trait's two lowest values. R/qtl2 drops H on a riself panel (R/qtl's convert2riself), so it keeps 173 animals there but only 2 of one class: nearly monomorphic, which is why its scan peaks elsewhere. GEMMA keeps the 20 hets at dosage 0.5, and that is where its sharper p-value comes from. GEMMA's next two hits are the same pattern in opposite directions (one A among 171, then two B among 171), so the chr16 block is a haplotype carried by a handful of extreme animals, not a population effect. Also fixes the reading of GEMMA's `af`: it is mean(dosage)/2 under the BIMBAM coding this deployment ships (B=1, H=0.5, A=0), so 0.469 for JAX00418132 is half the B-allele frequency 0.935, not the A-allele frequency 0.065. The bimbam file and DOL.geno agree at that marker. tests/frequencies.rs gains a route test: the class fields, A < B, the small-class note, the "not served" note, no individual names anywhere in the payload, and a 404 for a trait the group does not have.
8 daysgenes: Wikidata aliases, cached in gene-aliases.lmdb for seven daysPjotr
A GeneNetwork trait page shows the gene's aliases next to the symbol, and GN4 had no way to ask for them: gn-guile has a Wikidata lookup (gn/db/sources/wikidata.scm) and nothing in GN4 did. This is that lookup with the caching Wikidata deserves -- it is a shared public service and an alias list is not news. * GET /genes/{symbol}/aliases: two SPARQL queries against WIKIDATA_ENDPOINT (the entity: a gene wdt:P31 wd:Q7187 in human, mouse or rat, labelled with the symbol; then its skos:altLabel s in English), cached in WIKIDATA_CACHE_DIR/gene-aliases.lmdb, an LMDB file (NO_SUB_DIR, 0600, so one file and LMDB's -lock companion rather than a directory of two). * TTL WIKIDATA_TTL_SECONDS, seven days by default. Past it the value is refetched; ?refresh=true forces that. * The label is matched case-insensitively, which the gn-guile query did not: human genes are labelled SHH and mouse ones Shh, so an exact match finds only one species. ?species=human now returns the ortholog's labels -- HLP3, HPE3 and the rest -- which are the ones a mouse trait page shows. With no species named the pick is deterministic (mouse, rat, human), so a repeated request cannot change entity. * A failed refresh keeps the last value: the answer says stale: true, the envelope is partial, and a note explains it. Nothing cached at all is a 502 that says so. An empty alias list and an unreachable upstream are different facts, and only one of them is about the gene. Seven unit tests (key scoping, TTL including a clock that went backwards, the query shape and quote-escaping, SPARQL flattening, an LMDB round-trip that asserts the file is gene-aliases.lmdb, and stale-on-failure with and without a cache) and two route tests (the second call is served from the cache -- same fetched_at -- and the species switch), which skip with a 502 assertion when there is no network.
8 daysprobesets: cross-dataset lookups by probeset id and by gene symbolPjotr
A probeset id is not a Xapian field, so /search cannot find 1427571_at exactly and the only route to it was guessing a dataset and calling /datasets/{ds}/probesets/{name} until one answered. A gene symbol could be searched, but the reply was capped at 100 hits, interleaved probesets and platforms, and carried no total -- so "did I see all of Shh?" had no answer. Both questions are one indexed query each: ProbeSet.Name_IDX and symbol_IDX. * GET /probesets/{name}: one row per dataset that carries the probeset, with the panel (group, species, platform) and the scan's statistics (mean, LRS, LOD, locus, additive). Suffixed chip variants match too -- 1427571_at also finds 1427571_at_A -- and each row's `name` says which matched, so the answer never hides a substitution. LIKE's wildcards are escaped, or every Affymetrix `_at` would match any character. 155 rows over 155 datasets in 32 ms. * GET /genes/{symbol}/traits: the same rows filtered on Symbol, with a real total, ?sort=lod, ?min_lod=, ?group=, ?species=, ?platform= and paging -- "BXD datasets where Shh is a strong eQTL" is one call. 1225 rows over 17 datasets and 13 probesets in 38 ms. * Both serve metadata, not values. A token is enough for the index; trait values stay per-dataset and authorized per dataset, and `next` goes to /datasets/{ds}/probesets/{name}/values, which checks. There is no cross-dataset values endpoint, deliberately: that is bulk work and /groups/{g}/traits is the bulk call. The one-statement join took 5.2 s for Shh -- MySQL rescans InbredSet/Species per row -- so the 933 ProbeSetFreeze rows and the 102 panels are fetched separately and joined in memory: 70 ms each. nsamples is left out: it is a ProbeSetData count per row (+0.5 s) and the dataset record already reports strains.
8 daysapi: every response suggests where to go nextPjotr
A caller that knows only /version should be able to reach anything by following links. Until now `next` was thin and uneven: the genotype subtree offered none at all, the group record offered one, /version offered nothing, and two links it did offer could not be followed. * suggest:: now builds the graph: root_links (groups, species, mapping-methods, openapi), genotype_version_links (frequencies, marker names, a marker's location and tally, the group, the scanning actions), unlisted_genotype_links, marker_links (location, tally, marker list, panel tally, PheWAS at that marker), group_links and trait_links. Envelope::with_links / with_links_except add them, deduplicating and never suggesting the response the caller is already reading. * Wired through the genotype surface (listing, version detail, markers, location), both frequencies routes, the compute endpoints, groups, traits, both value routes, species, datasets, publications, accessions, search, mapping methods, phewas, correlation and /version. * Fixed what was not followable, because `next` promises a GET that returns 200 and a client is the one that finds out otherwise: - /groups/{g}/phewas without a locus is a 422. Bare PheWAS links are gone; comput endpoints now link ?locus=<their strongest marker> (suggest::peak_marker_of, computed before the rows move). - /datasets and /publications are not routes; root_links offers the listings that exist (/groups, /species, /mapping-methods) and /openapi.json. - /search?q= with an empty query is a 400, so it is not a link. - the correlation endpoint's "/compute/correlation" was a POST in `next`; it is in `actions` now. - marker_links pointed at /markers/{marker}, which is not a route; it points at /markers/{marker}/location. tests/links.rs walks the graph two hops from a dozen seeds and fails on any link that does not answer 200, plus a check that a tally leads to the marker list and to a marker's location, and a location to that marker's tally and its PheWAS.
8 daysgenotypes: call frequencies, marker names, and a location per markerPjotr
Two questions the genotype surface could not answer: what a panel is actually called as, and which markers it holds. The listing gives counts and the compute endpoints give scans; a panel whose calls are not the codes its header declares, or a marker that is 99% missing, stayed invisible until a scan came back flat -- and then the search for the cause started at the wrong end. * GET /groups/{group}/genotypes/{version}/frequencies counts the calls in the version's .geno and returns them as call -> frequency ({"B": 0.4889, "D": 0.492109, "H": 0.018991}). Frequencies are over the calls that are present, so they sum to 1; missing codes (U, N, NA, an empty cell) are counted separately and reported in notes, because the missing rate is the diagnosis. * A probability-valued file is binned 0.05 wide (twenty bins over [0, 1], keys are the bin edges), since 0.4187 and 0.4191 are the same statement about a marker; ?bin= sets another width. A file with a few distinct numbers (0, 0.5, 1 for a biallelic dosage) is reported exactly -- binning three values hides what is being looked at. * .../markers/{marker}/frequencies is one locus, and ?marker= on the panel route is the same call. Counting goes through an interner, so cost is one hash lookup per cell and one allocation per distinct call; a .geno over 256 MB is refused (413 genotype_too_large) rather than read, which is what HSNIH-Palmer_true (3.7 GB, 134927 markers x 6147 individuals) demanded. * The version's /markers now lists marker NAMES only, one column, and .../markers/{marker}/location returns {"name","chr","cM","Mb"} as named fields: a page of names stays small, and a position is a lookup rather than a table the client must hold. The same size guard applies, so an oversized panel's listing is a 413 in 3 ms. * Files no group lists -- a panel in preparation, a fixture, the DO's do/ cross -- live at /groups/unlisted/genotypes/{name}/... with the same four calls. `unlisted` is a namespace, not a group: only its genotype subtree answers. One prefix for the whole genotype surface beats a second spelling for the same tally, so the earlier /genotypes/{name} spelling is gone. * genotypes::resolve_named (moved out of the frequencies module, since it is genotype metadata) resolves a version, a .geno name, or a path relative to the genotype files directory; MAX_SCAN_BYTES is shared by every whole-file read. Tests limited to the feature: genotypes 3 (names-only columns, paging, location fields, marker_not_found, unlisted parity, 413), frequencies 5, frequencies unit 7.
8 dayscompute/qtl2: multiparent crosses, by namePjotr
A multiparent population cannot come from a .geno. That file carries two codes per marker (B/D, A/H/B) -- a collapse, not a founder allele -- and the collapse is not invertible, so DOL's 194 DO animals cannot be fitted as a DO however much reference data appears. What can be is a cross that ships its own qtl2 control file with founder genotypes, so the endpoint now takes one by name: {"rqtl2_cross": "do", "trait_name": "DO::10001 (WBC)", "samples": ...} * resolve_control_file looks in genotype/rqtl2/{name}.json, then {name}/{name}.json (do/do.json), then {name}.json under GENOTYPE_FILES; prepare_control_cross reads the cross's pheno, maps and founder_geno headers without converting anything. It reports crosstype, founders and whether the cross is multiparent, and refuses a multiparent crosstype with no founder_geno rather than silently scanning it as biallelic. * Samples are matched to the cross's own population by name and the strays are counted in notes; rqtl2_cross cannot be combined with group or genotype_file. * The gate is given individuals x founders, because a multiparent scan fits one coefficient per founder allele at every marker. * Autosomes only now applies to shipped crosses as well: the DO's X is left out (369 markers) instead of being fitted without a sex covariate, and the script subsets with cross[, chr = autosomal]. Measured on the DO (WBC, 742 individuals, 8 founders, 6044 autosomal markers): probabilities 174 s / 1.39 GB (cached), LOCO kinship 15 s (cached), scan 21 s. scan1 runs on one core for a multiparent cross because its forked path stalls on allele probabilities here -- 25+ minutes at 8 cores with no CPU consumed, against 21 s serial. Two bugs found by running an actual published trait through it: * do_pmap.csv is in Mb, not the bp qtl2 documents, so every DO peak was reported at 0.0 Mb. The map's unit is now detected from its largest position and printed. * The phenotype file was read with the default whitespace separator, so a label with a space ("DO::10001 (WBC)") split into three fields and read.table stopped with "line 1 did not have 3 elements". Both the pheno and covariate reads are tab-separated now. The R error extraction also joins R's wrapped continuation line, so a complaint reads as a sentence rather than "Error in scan(...) :".
9 dayscompute: R/qtl2 as a third scan, POST /compute/qtl2Pjotr
R/qtl sits between the existing two scans: Haley-Knott on marker genotypes, no kinship. GEMMA is an LMM but biallelic BIMBAM with mean-imputed genotypes. Neither can analyse a multiparent population, and neither models relatedness from the actual genotype data. R/qtl2 does both -- HMM genotype probabilities, then an LMM with leave-one-chromosome-out kinship -- and it is what names the third peak on DOL trait 10003 that R/qtl's scan cannot see (chr16, LOD 4.24, against GEMMA's -log10 p 7.15 there). * src/qtl2.rs converts this deployment's own .geno into a qtl2 cross under QTL2_CACHE_DIR: control.json plus geno/gmap/pmap/pheno, with the genotype-code map, the orientation and the paths written out explicitly. The shipped genotype/rqtl2/ control files cannot be used -- they point outside the data directory, declare no code map (every genotype reads as missing) and no orientation (individuals read as markers) -- and each of those produces a plausible flat scan instead of an error, which is why the conversion is ours. * Heterozygous calls are dropped on a riself panel, as R/qtl's convert2riself does. Left in, qtl2 reads the cross as multiparent and its scan1 takes a path the panel cannot satisfy ("Error in scan1_pg ... : No individuals in common"). * Autosomes only: X needs a sex covariate to be fitted properly and Y/MT degenerate the per-chromosome kinship. GEMMA's LOCO scan is restricted the same way. The note says how many markers were left out. * scripts/qtl2_scan.R caches genotype probabilities and kinship by a hash of the control file *and its directory*; two conversions of the same panel can have identical control files and identical file sizes, and reusing the first one's probabilities silently scans the wrong cross. It also takes the individual names from whichever shape the cross's pheno slot has -- a single-column pheno file leaves it a plain matrix, and the resulting empty phenotype reads as scan1's "No individuals in common". * Rows are marker, chr, pos, lod, so /plot/rqtl draws the reply unchanged -- post the same samples as /compute/rqtl. * tests/qtl2.rs: the fixture scans and its notes name the cross; BXD trait 10001 peaks on chr8 under both scans; refusals carry missing_genotype / invalid_kinship. cost: 2 concurrent scans, gated from RQTL_LARGE_PANEL individuals like R/qtl; a DO cross costs minutes of genotype probabilities, paid once per panel. QTL2_SCRIPT_PATH and QTL2_CACHE_DIR configure it.
9 daysdocs: the skill says fitqtl fits covariatesPjotr
`/compute/fitqtl` has taken `covariates` since c642f87, and the skill mentioned it only as a trailing sentence inside the fitqtl bullet -- after the paragraph that says "both engines", which an assistant skimming for covariates reads as rqtl-and-gemma. Three changes so the offer is where a reader looks for it: * the covariates bullet now names all three endpoints (`/compute/rqtl`, `/compute/gemma` **and `/compute/fitqtl`**); * the fitqtl bullet says it in bold, with why it matters (the `additive` is then the effect *after* conditioning), and adds the consequence that the LOD no longer matches the scan once a covariate is fitted -- the earlier wording claimed parity unconditionally; * the endpoint table gained `POST /compute/{rqtl,gemma,fitqtl}` (it had no entry at all), and the session outline now includes sizing a peak with fitqtl when the user asks how much it matters. Docs only.
9 dayscompute: fit covariates in /compute/fitqtl tooPjotr
`fitqtl` takes a `covar` argument, so the effect it reports can be the one *after* conditioning -- which is usually the number worth quoting for a peak that overlaps a known confounder. `covariates` now works on /compute/fitqtl exactly as on /compute/rqtl: the same `covariates::resolve` (case attribute by name, inline values, explicit map, same coding and same note), so the two endpoints report a covariate identically. The R side does what R/qtl needs: the phenotype file carries one column per covariate after the trait, `--covarstruct` says which are categorical, and the script dummy-expands those (k-1 against the first sorted level, missing values left missing). Two things it has to get right: * `model.matrix()` drops rows with a missing value, which silently misaligns the covariate frame against the cross -- so the dummies are built directly (`ifelse(is.na(x), NA, x == level)`), not through `model.matrix`; * a covariate can be perfectly spread across the panel and **constant among the individuals a trait covers** -- BXD's `Status` is exactly that -- and R/qtl then stops with `Lapack routine dgesv: system is exactly singular`. Columns that do not vary among the phenotyped individuals are left out, and the reply names them, the same rule the GEMMA covariate file applies. Measured on BXD trait 10001 at its peak, `BXD.4`: no covariate LOD 2.9310 additive +2.394 32.8% var with Status LOD 3.3602 additive +2.463 36.6% var with an inline numeric covariate LOD 2.9373 additive +2.361 The LOD is no longer the scan's once a covariate is fitted -- it is a different model -- which the docs now say, next to the no-covariate parity check that still holds. * scripts/qtl_fitqtl.R: `--covarstruct`, dummy expansion, the degenerate column rule, and the dropped names in the summary line. * src/fitqtl.rs: writes the covariates into the phenotype file and the struct file, and passes both. * src/routes/fitqtl.rs: the `covariates` field, resolved after alignment so the coding and the complete-case count match /compute/rqtl. * tests/fitqtl.rs: a covariate run (the note, the fitted row, a changed LOD) and an unknown attribute refused with the list of valid ones. * README.md, docs/www/SKILL.md: covariates on this endpoint, and why its LOD stops matching the scan's. 306 tests pass; the same four as before do not.
9 dayscompute: add /compute/fitqtl for effect sizes at named markersPjotr
`/compute/rqtl` answers "where" -- scanone gives a LOD per marker -- but not "how much": a scan statistic is not a fitted model. R/qtl's `fitqtl` fits one, and reports the QTL's effect, its standard error, the model LOD and the variance explained. That is what a table of effects needs, and it is a separate endpoint because it is a separate question, asked about a handful of markers (a scan's peaks) rather than a genome: `fitqtl` fits one model per marker, so a genome-wide version would be thousands of fits. Cap: 100 markers per request. Rows are positional: marker, chr, pos, lod, pct_var, additive, se, t, p `additive` is fitqtl's estimate for the QTL term -- R/qtl codes the QTL genotype +-1, so it is half the difference between the two homozygote classes, the same convention GeneNetwork's published `additive` uses. **Two independent checks that the numbers are right.** On BXD trait 10001 (`BXD.4`), at the scan's peak: * `lod` = 2.930985, and `/compute/rqtl`'s scanone gives 2.930985 at the same marker -- identical to every digit the CSV carries, because they are the same cross and the same phenotype. That parity is now an integration test; * `additive` = 2.394472, against the value genenetwork.org publishes for that trait (`/api/v_pre1/trait/BXD/10001` -> 2.394444), which is 28 ppm away -- and that published number came from qtlreaper, a different engine. * scripts/qtl_fitqtl.R (new): reads the same `.geno` and two-column phenotype file as `rqtl_wrapper.R` (which is a verbatim copy of GN3's script and so not extended), builds the cross the same way, fits `fitqtl` at each named marker with `get.ests=TRUE`, and writes one CSV row per marker -- plus a one-line summary naming the strongest fit. `calc.genoprob(step=0)`: marker-level probabilities are all a marker-level fit needs. * src/fitqtl.rs (new): runs it, parses the CSV into positional rows (`chr` stays text, it can be `X`), keeps the command line for `notes`. * src/routes/fitqtl.rs (new): the endpoint -- same alignment rules as `/compute/rqtl` (by name onto the panel, parents/F1s excluded unless asked), the same `normalize`/`winsor_*` pipeline (effects come out on that scale), and the same R/qtl gate, so a fitqtl request and a scan queue against one another rather than competing for cores. * src/config.rs: `FITQTL_SCRIPT_PATH` (default `scripts/qtl_fitqtl.R`). * tests/fitqtl.rs: the fixture (columns, sign of the effect, notes), the LOD parity against scanone, and the refusals (no markers, both genotype inputs). * README.md, docs/www/SKILL.md: the endpoint, what each column means, and when to reach for it instead of the scan. 305 tests pass; the same four as before do not.
9 dayscovariates: refuse a mostly-numeric attribute, and teach the skill to use themPjotr
Fitting `SeqCvge` as a covariate came back as `502 gemma_failed` with a message about LU decomposition: its values are depths (`40`, `41`) with a few two-library entries (`36+39L`), so the attribute was coded as **twenty categories** and GEMMA died on the resulting dummies. The failure named GEMMA, which is the wrong place to look. `code_case_attribute` now refuses that shape at the covariate step, naming the values and what to send instead: 422 invalid_covariate: case attribute 'SeqCvge': the attribute mixes numbers with number-like text (14 of 19 values look numeric, but e.g. '36+30L' does not), so coding it as categories would fit a quantity as a factor -- and GEMMA fails on the result. Pass an explicit `map` giving each of those a number (there are 5 such value(s): ...), or send the numbers inline. The rule is deliberately narrow: an odd level with *no* digits (`1, 2, unknown`) is still a level and still coded, so nothing that worked stops working; only dirty numbers are caught. Also, the skill's covariate coverage -- which the review above found thin: * when a covariate is worth trying (metadata that could confound: sex, age, epoch, batch, sequencing depth), and the covariate step in the session outline; * read `.data.descriptions` before choosing one -- they are the only place that says `SeqCvge` is sequencing depth and `Epoch` a family subgroup -- and the endpoint table now says so; * read the covariate line in `notes` for the coding and the complete-case n, and report that n beside any LOD/p change; plus `Cyro`/`Cryo` being two levels in BXD; * the categorical-vs-numeric rule measured rather than asserted: `RRID` refused (82 distinct), `Status`/`Epoch` accepted, and the compound-value trap written up with the `map` workaround; * GEMMA's dummy columns, the column it drops when a level has no fitted individual, and that R/qtl's note has no such line; missing values dropping individuals; the 8-covariate cap and the `group` requirement. * src/covariates.rs: the refusal, with a unit test for both sides of it; * tests/gemma.rs: `SeqCvge` is a 422 naming `map`, not a GEMMA failure; * docs/www/SKILL.md: the covariate section, the session line, and the case-attributes rows in the endpoint table. 300 tests pass; the same four as before do not.
9 daysapi: normalize trait values, and winsorise the tails, before mappingPjotr
GeneNetwork2's trait page offers a set of normalisations, and its mapping page warns about outliers without ever fixing them. This adds both, under one spelling that `/compute/rqtl`, `/compute/gemma` and `/groups/{g}/traits/{id}/values` all share: normalize none | log2 | log10 | sqrt | negate | zscore | robust_zscore | rank (GN2's names `invert` and `qnorm` accepted) winsor_quantile clip each tail at a fraction, e.g. 0.05 winsor_fences clip at Q1-k*IQR / Q3+k*IQR, GN2's outlier rule They run in a fixed, reported order -- transform, then winsorise, then standardise -- so clipping acts on the scale being analysed and the result lands exactly on mean 0 / sd 1 rather than on the sd of clipped data. The note gives the steps *in the order they ran* (`winsorized at Tukey fences, then zscore`), the count clipped in each tail, and how many values sit outside Tukey fences -- the warning GN2 gives before mapping and this API gave nowhere. **Checked against genenetwork.org itself.** GN2's public trait page embeds its own reference values (`js_data.qnorm_values` / `zscore_values`, the `data-qnorm`/`data-zscore` attributes), so there is an oracle for the two transforms whose definition matters. On BXD trait 10010 (28 samples): `rank` matches to 0.0005 and `zscore` to 0.0005 -- i.e. GN2's three-decimal rounding. Getting there found a real disagreement: GN2 uses `scipy.stats.mstats.zscore`, whose default is `ddof=0`, so its z-scores use the **population** sd; ours used the sample sd and came out `sqrt((n-1)/n)` low (1.8% at n = 28). We now match GN2. Two GN2 quirks deliberately not reproduced: its "Quantile" is a rank-based inverse-normal (Blom) transform, so it is called `rank` here; and its `sqrt` silently adds 1 when a value is in [0,1], which shifts every effect size -- a negative value is refused instead. * src/normalize.rs (new): the transforms, percentile/Tukey fences, Blom scores over an Acklam inverse-normal, `Params` for query/body use, `to_json` for suggestions, and eleven unit tests -- the quantile table, ties sharing an average rank, missing values never imputed, GN2's own `find_outliers` example, and the note's wording. * src/routes/{rqtl,gemma}.rs: the three fields, applied **after** alignment (over the samples that enter the fit) and before anything reads the values, with the note; the R/qtl permutation suggestion and GEMMA's R/qtl suggestion carry the pipeline. * src/routes/groups.rs, src/suggest.rs: `?normalize=` on the values endpoint, and the mapping bodies in `actions` carry the same fields -- a caller who fetches z-scores must map those numbers, not the raw ones. * tests/groups.rs: values are mean 0 / population sd 1, the note says so, every `/compute/*` suggestion in `actions` carries it, and an impossible pipeline is a 422 `invalid_normalization`. * README.md, docs/www/SKILL.md: what each name means, the order, the population-sd parity, and that effects come out on the transformed scale. 298 tests pass; the same four as before do not.
9 dayscompute: report GEMMA's effect size next to the scorePjotr
A GEMMA row now carries the effect beside the p-value it belongs to: `… beta, se, l_mle, additive, p_lrt`. `additive` is `-beta/2`, which is what GeneNetwork's own tables and CSV call the additive effect (gn2/wqflask/marker_regression/gemma_mapping.py:240). Halved because half the difference between the two homozygotes is the usual additive effect; negated because GN2's BIMBAM conversion codes `@mat` as 1, so GEMMA's beta is per *maternal* copy while GeneNetwork's published sign is anchored on the other allele. Two facts that make this worth doing carefully: * `beta` and `se` are left exactly where GEMMA put them. `se` is what makes the effect interpretable (`beta +/- 1.96*se` is a 95% interval) and both are GEMMA's own numbers, so replacing them with a derived one would lose information; * the derived column goes *before* the p-value, not after it. `p_lrt` must stay the last column: `gemma::peak_marker` reads `row.last()` for the peak (and with it the PheWAS and plot suggestions), and the plotters and the docs all say "the last column is the p-value". Appending would have silently pointed every peak at an effect size. The response also says what the three columns mean, and that they describe *this run* on these samples and this genotype version -- not the `additive` trait info reports, which came from the published mapping (trait 20010: this run -0.54, published +2.118, same peak marker). Nothing is added when GEMMA wrote no `beta` (a non-LMM run): the columns are then exactly GEMMA's. * src/gemma.rs: `with_effect_size`, applied by `parse_assoc_file`, so /compute/gemma and /compute/gemma/ravanan both get it. * src/routes/gemma.rs: the note. * tests: four unit tests (placement, sign and halving, `peak_marker` still finds the peak from the last column, a table without beta untouched, a null effect for a row with `NA` beta) and one end-to-end test asserting the column sits immediately before `p_lrt`, every value is `-beta/2`, `p_lrt` stays a probability, and the note is there. * README.md (the response example and what the column means), docs/www/SKILL.md (how to read and quote it). 286 tests pass; the same four as before do not.
9 dayscompute: let GEMMA name the trait, so its plot can be titledPjotr
Two GEMMA runs over one BXD panel drew genuinely different pictures -- traces differ on all 20 chromosomes, peaks -log10 p 5.48 at chr8 97.0 Mb against 5.43 at chr1 175.6 Mb -- and both came out headed **"trait gemma: GEMMA association along the genome"**, because GEMMA's request had no trait label at all: `/compute/gemma` fell back to `<group>::gemma`, and `/plot/gemma` fell back to "gemma". The plot of each trait was told apart by its annotation and nothing else. `trait_name` is now an optional field on `GemmaRequest`, used for the `Trait: ...` note (first in `notes`, as on `/compute/rqtl`), for the PheWAS and R/qtl suggestions, and for the `/plot/gemma?trait=` suggestion -- so a posted response titles its own plot, and `?trait=` still overrides. Without it nothing changes: the label stays `<group>::gemma`. * src/routes/gemma.rs: the field, the label resolution, the note, the plot href. * tests/gemma.rs: the covariate run names its trait and asserts the note is first; tests/plot.rs: a GEMMA plot titles itself from the note, and an explicit `?trait=` wins over it. * README.md (`/compute/gemma`, and the plot section), docs/www/SKILL.md (the GEMMA options list, and what the plot title comes from). 281 tests pass; the same four as before do not.
9 daysSKILLpjotr