Every change Wellknown observed on this MCP server, newest first, with what it was before and what it became. Tool-surface changes carry the definition diff. Nothing here is edited after the fact.
Changed the definition of "fetch_gene_for_expression", "list_models" and "predict_expression"
Fetch a gene's sequence prepared for expression prediction. Resolves the gene's canonical-transcript TSS via Ensembl andreturnsstores a9,198 bpgene-sensewindow with the TSS at offsetsequence4,599centred(itsonmidpoint)it, as ahandle to pass topredict_expression(sequence_ref=...).ThatLongerisTSS-centredthesequenceminimum validgivesinputbetterforpredictions,everysoexpressionthismodel,alreadyandfetchesatas9,198muchbpflankitonneedseachnoside`tss_index`. ItofistheallTSS as thedefaultlistedg0-expressionexpressionreads,modelsbutrecommenda(the largestmodelbio_spec.recommended_flank_bpmayfromrecommendlist_models). One handleaserveswidereveryflankexpression model:bio_spec.recommended_flank_bpthefromAPIlist_modelsreads(e.g.what the chosen model40,960needsbparoundfortheg0-expression-8192).TSSToandsupplyignoresflanktheF,rest. Thecall fetch_regionhandle records `tss_index` (region=<thisthehandle'sTSSregion>,offsetstrand=<itsintostrand>it), andflank_bp=Fpredict_expression-uses4,599)it when you do not pass one,whichsoreturnsno2Foffset arithmetic is needed. Every expression model accepts at least 9,198 bp with the TSS atoffsetleastF4,and599passbptss_index=Ffromtoeachpredict_expression.end.IfNear a chromosome end thereturnedflanksequenceisnotshortenedexactlyto2Fwhatbpfits (the gene isreportednearas `flank_shortened`),aneverchromosomebelowend)4,599thebp; a TSSis notcloserattoF:theuseend thanathatsmallerisflank.an error.
⟨99 unchanged words⟩ field where there is no sliding window (9,198 for g0-expression;null foran expressiona modelwithout one, such as g0-expression-8192). For expression, `recommended_flank_bp` is how manybptothe modelsupplyreads on each side of theTSSTSS;(4,599longerforTSS-centredg0-expression,input40is better,960
⟨50 unchanged words⟩ TSS at least 4,599 bp from eachend. Longer input is welcomeend, up to500,000 bp.EachLongermodel'sTSS-centred sequence gives better predictions: each model's bio_spec.recommended_flank_bp (see list_models) says howmany bpto supplyit reads on each side of the TSS: 4,599for the default g0-expression, which reads 9,198 bp around the TSS,and40,960 for g0-expression-8192, whichflankreadsbeyondmuch
Changed the definition of "find_genes" and "find_genes_and_predict_expression"
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":truefalse}
{"destructiveHint":false,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":truefalse}
Changed the definition of "fetch_gene_for_expression", "find_genes", "find_genes_and_predict_expression" and 2 more
⟨8 unchanged words⟩ Resolves the gene's TSS via Ensembl and returnsthe exact TSS-centreda 9,198 bp gene-sense window with theexpressionTSSmodelatscoresoffset 4,599 (its midpoint), as ahandlehandle to pass to predict_expression(sequence_ref=...).BecauseThat
Certificate recorded, valid to 2026-12-17
Authorization not required
First tool surface recorded: 15 tools (server version 0.1.0a20)
Showing the latest 6 events. The API returns up to 500 and filters by kind: ?kind=tool_surface_changed
⟨36 unchanged words⟩ ,198 bp is higher, and that isathefixedsequencewindowaroundratherone TSS rather than aminimumregionsize.to search.) Gene-finding: detects transcript boundaries (TSS + ⟨146 unchanged words⟩
⟨63 unchanged words⟩ predict their expression in K562". predict_expression scores the sequence around ONE TSSwindowand needs you toknow wherethat TSS is (eitherapre-centred9,198 bp window with the TSS at its midpoint, or a longer sequence plus `tss_index`);this tool discoversevery gene's TSS itself. It has no 9,198 bp floorand no tss_index;it starts with gene finding, so it takes1,000–500,000 bp. Runs async internally at ⟨64 unchanged words⟩
⟨102 unchanged words⟩ is no sliding window (9,198 forg0-expressiong0-expression; null for an expression model without one, such as g0-expression-8192).`request_max_bp`For expression, `recommended_flank_bp` isthehowonlymany bponeto supply on each side of thethreeTSSthat(4,599isforag0-expression, 40,960cap;for g0-expression-8192). `request_max_bp` is thewindowonly cap; the other fields describe what the modelscoresreads, not what theroute accepts.
Predict a gene's expression fromatheTSS-centredsequencewindow.around its TSS. Expression is cell-type-specific, so `description` (cell type ⟨9 unchanged words⟩ — the API rejects requests without it. Themodelinputscoresruleexactlyis the same for every expression model: at least 9,198 bpcentred,onwith the TSS(±4at least 4,599).Twobp from each end. Longer inputwaysis welcome, up to 500,000 bp. Each model's bio_spec.recommended_flank_bp (see list_models) says how many bp to supplythat: -onAeachsequenceside ofexactlythe TSS: 4,599 for the default g0-expression, which reads 9,198 bpalreadyaroundcentredtheonTSS,theandTSS.40,960No `tss_index`for g0-expression-8192,neededwhich—readsthemuchmidpointfurther.isLesstheflankonlythanlegalaTSS.model-recommends,Adownlongertolocus4,5999bp,198–500is accepted and scored,000but that model then reads less than it was trained on. - A 9,198 bp sequence with the TSS at offset 4,599plus(the midpoint) needs no `tss_index`. - Any longer sequence needs `tss_index`: the 0-basedoffset of theTSS into it. The API cutsthewhatwindowtheformodelyou (sequence[tss_index-4599reads:aroundtss_index+4599])that offset and never scans for aTSS itself.Anything underUnder 9,198 bp,isorrejecteda TSS closer than 4,599 bp to either end, is rejected here and by the API (422).— thereAislocus with nopaddingoffsetoris rejectedtruncationtoo,fallback.because`tss_index`it isrequiredindistinguishable from a mis-centredforwindow.everyAotherverylength,N-richbecauseoralow-complexitylocuswindowwithcannoalsooffsetbeisrefusedindistinguishable frombyathemis-centredAPIwindow.(422). An offset that is merely WRONG (e.g. ⟨30 unchanged words⟩ response. Easiest paths: fetch_gene_for_expression(gene) returns aready-centred9,198 bp handle with the TSS at its midpoint, valid for every expression model, and find_genes_and_predict_expression takes araw region andfinds each TSS for you.
⟨179 unchanged words⟩ sequence is exactly 9,198 bp; must leave at least 4,599 bp on eachside.side (4,599 <= tss_index <= length - 4,599). TheAPImodelscoresreadsonlythe sequence[tss_index-4599:aroundtss_index+4599]this offset (9,198 bp for g0-expression, more flank for a model with a larger bio_spec.recommended_flank_bp) and the API reports the slice it used asmeta.task_specific_counts.scored_window —meta.task_specific_counts.scored_window.checkCheck it: a wrong-but-in-range offset scores the wrong ⟨9 unchanged words⟩