jbrowse-plugin-protein3d 0.5.1 → 0.5.3

package/README.md

@@ -11,6 +11,16 @@ It has features to automatically look up a protein structure of interest using
 the UniProt ID mapping API to connect to AlphaFoldDB, and can also use Foldseek
 to look up related structures also
 
+## Coordinate-mapping harness
+
+A standalone diagnostic page that loads real PDB / AlphaFold structures through
+the plugin's actual mapping code and surfaces cases it mishandles (multi-chain
+complexes, partial/repeat structures, AlphaFold fragments):
+
+https://gmod.org/jbrowse-plugin-protein3d/
+
+Source and details in [harness/](harness/).
+
 ## Screenshot
 
 ![](img/1.png)

package/dist/AddHighlightModel/ProteinToMsaHoverSync.js

@@ -2,6 +2,7 @@ import { useEffect } from 'react';
 import { getSession } from '@jbrowse/core/util';
 import { autorun, untracked } from 'mobx';
 import { observer } from 'mobx-react';
+import { findConnectedMsaView } from './findConnectedMsaView';
 import { findStructureRowName } from './msaRowMatch';
 import { getProteinView } from './util';
 import { stripStopCodon } from '../LaunchProteinView/utils/util';
@@ -9,10 +10,13 @@ const ProteinToMsaHoverSync = observer(function ProteinToMsaHoverSync({ model, }
     const session = getSession(model);
     const { views } = session;
     const proteinView = getProteinView(session);
-    const connectedMsaViewId = proteinView?.connectedMsaViewId;
-    const msaView = connectedMsaViewId
-        ? views.find(f => f.id === connectedMsaViewId)
-        : undefined;
+    // pair with the MSA either by an explicit connectedMsaViewId or, in the
+    // genome-centric flow, by the genome view this structure and the MSA both
+    // connect to (see findConnectedMsaView)
+    const msaView = findConnectedMsaView(views, {
+        connectedMsaViewId: proteinView?.connectedMsaViewId,
+        structureViewId: proteinView?.primaryStructure?.connectedViewId,
+    });
     useEffect(() => {
         if (!proteinView || !msaView) {
             return;

package/dist/AddHighlightModel/findConnectedMsaView.d.ts

@@ -0,0 +1,23 @@
+interface MsaViewLike {
+    id: string;
+    type: string;
+    connectedViewId?: string;
+}
+/**
+ * Find the MsaView a protein view should sync hover with. Two ways they pair:
+ *  - an explicit `connectedMsaViewId` (set when the protein view was launched
+ *    from an alignment), or
+ *  - a shared genome view: the protein structure and an MsaView are both
+ *    connected to the same LinearGenomeView via `connectedViewId` — the
+ *    genome-centric gene-explorer flow, where no explicit MSA link is set but
+ *    both views already bridge through the same genome coordinates.
+ *
+ * Mirrors react-msaview's structureMatchesMsa (the MSA→structure side): a shared
+ * genome view is sufficient to connect, so neither side has to thread an
+ * explicit cross-view id.
+ */
+export declare function findConnectedMsaView<T extends MsaViewLike>(views: T[], { connectedMsaViewId, structureViewId, }: {
+    connectedMsaViewId?: string;
+    structureViewId?: string;
+}): T | undefined;
+export {};

package/dist/AddHighlightModel/findConnectedMsaView.js

@@ -0,0 +1,23 @@
+/**
+ * Find the MsaView a protein view should sync hover with. Two ways they pair:
+ *  - an explicit `connectedMsaViewId` (set when the protein view was launched
+ *    from an alignment), or
+ *  - a shared genome view: the protein structure and an MsaView are both
+ *    connected to the same LinearGenomeView via `connectedViewId` — the
+ *    genome-centric gene-explorer flow, where no explicit MSA link is set but
+ *    both views already bridge through the same genome coordinates.
+ *
+ * Mirrors react-msaview's structureMatchesMsa (the MSA→structure side): a shared
+ * genome view is sufficient to connect, so neither side has to thread an
+ * explicit cross-view id.
+ */
+export function findConnectedMsaView(views, { connectedMsaViewId, structureViewId, }) {
+    const msaViews = views.filter(v => v.type === 'MsaView');
+    const byExplicitId = connectedMsaViewId
+        ? msaViews.find(v => v.id === connectedMsaViewId)
+        : undefined;
+    const bySharedGenomeView = structureViewId
+        ? msaViews.find(v => v.connectedViewId === structureViewId)
+        : undefined;
+    return byExplicitId ?? bySharedGenomeView;
+}

package/dist/LaunchProteinView/components/AlphaFoldDBSearch.js

@@ -56,11 +56,11 @@ const AlphaFoldDBSearch = observer(function AlphaFoldDBSearch({ feature, session
                 React.createElement(ExternalLink, { href: "https://www.uniprot.org/" }, "UniProt"),
                 ' ',
                 "directly and use \"Enter manually\" above, or use \"Search sequence against AlphaFoldDB API\" if available.")),
-            state.showStructureSelectors && (React.createElement(React.Fragment, null,
+            state.showStructureSelectors && state.isoformSequences ? (React.createElement(React.Fragment, null,
                 React.createElement("div", { className: classes.selectorsRow },
                     React.createElement(TranscriptSelector, { val: state.userSelection, setVal: state.setUserSelection, structureSequence: state.structureSequence, feature: feature, isoforms: state.transcriptOptions, isoformSequences: state.isoformSequences })),
-                state.showSequenceSearchStatus && (React.createElement(SequenceSearchStatus, { isLoading: state.isSequenceSearchLoading, uniprotId: state.uniprotId, url: state.url, hasProteinSequence: !!state.userSelectedProteinSequence, sequenceSearchType: state.sequenceSearchType })),
-                state.showAlphaFoldDBSearchStatus && (React.createElement(AlphaFoldDBSearchStatus, { uniprotId: state.uniprotId, selectedTranscript: state.selectedTranscript, structureSequence: state.structureSequence, isoformSequences: state.isoformSequences, url: state.url }))))),
+                state.showSequenceSearchStatus && (React.createElement(SequenceSearchStatus, { isLoading: state.isSequenceSearchLoading, uniprotId: state.uniprotId, url: state.url, hasProteinSequence: !!state.userSelectedProteinSequence?.seq, sequenceSearchType: state.sequenceSearchType })),
+                state.showAlphaFoldDBSearchStatus && (React.createElement(AlphaFoldDBSearchStatus, { uniprotId: state.uniprotId, selectedTranscript: state.selectedTranscript, structureSequence: state.structureSequence, isoformSequences: state.isoformSequences, url: state.url })))) : null),
         React.createElement(DialogActions, null,
             React.createElement(ProteinViewActions, { handleClose: handleClose, uniprotId: state.uniprotId, userSelectedProteinSequence: state.userSelectedProteinSequence, selectedTranscript: state.selectedTranscript, url: state.url, confidenceUrl: state.confidenceUrl, feature: feature, view: view, session: session, alignmentAlgorithm: alignmentAlgorithm, onAlignmentAlgorithmChange: onAlignmentAlgorithmChange, sequencesMatch: state.sequencesMatch, isLoading: state.isLoading, error: state.error }))));
 });

package/dist/LaunchProteinView/components/AlphaFoldDBSearchStatus.d.ts

@@ -1,12 +1,10 @@
 import React from 'react';
+import type { IsoformSequences } from '../utils/util';
 import type { Feature } from '@jbrowse/core/util';
 export default function AlphaFoldDBSearchStatus({ uniprotId, selectedTranscript, structureSequence, isoformSequences, url, }: {
     uniprotId?: string;
     selectedTranscript?: Feature;
     structureSequence?: string;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
+    isoformSequences: IsoformSequences;
     url?: string;
 }): React.JSX.Element;

package/dist/LaunchProteinView/components/IdentifierSelector.js

@@ -1,46 +1,12 @@
 import React, { useState } from 'react';
 import { Button, FormControl, InputLabel, MenuItem, Select, } from '@mui/material';
-import { getDatabaseTypeForId } from '../utils/util';
-function getIdLabel(id) {
-    const dbType = getDatabaseTypeForId(id);
-    if (dbType === 'refseq') {
-        if (id.startsWith('NM_') || id.startsWith('XM_')) {
-            return `${id} (RefSeq mRNA)`;
-        }
-        if (id.startsWith('NR_') || id.startsWith('XR_')) {
-            return `${id} (RefSeq ncRNA)`;
-        }
-        if (id.startsWith('NP_') || id.startsWith('XP_')) {
-            return `${id} (RefSeq protein)`;
-        }
-        return `${id} (RefSeq)`;
-    }
-    if (dbType === 'ensembl') {
-        if (id.includes('G')) {
-            return `${id} (Ensembl gene)`;
-        }
-        if (id.includes('T')) {
-            return `${id} (Ensembl transcript)`;
-        }
-        if (id.includes('P')) {
-            return `${id} (Ensembl protein)`;
-        }
-        return `${id} (Ensembl)`;
-    }
-    if (dbType === 'hgnc') {
-        return `${id} (HGNC)`;
-    }
-    if (dbType === 'ccds') {
-        return `${id} (CCDS)`;
-    }
-    return id;
-}
+import { getDbIdLabel } from '../utils/util';
 export default function IdentifierSelector({ recognizedIds, geneName, selectedId, onSelectedIdChange, }) {
     const [expanded, setExpanded] = useState(false);
     // Build list of selectable options
     const options = [
         { value: 'auto', label: 'Auto (try all)' },
-        ...recognizedIds.map(id => ({ value: id, label: getIdLabel(id) })),
+        ...recognizedIds.map(id => ({ value: id, label: getDbIdLabel(id) })),
     ];
     if (geneName) {
         options.push({

package/dist/LaunchProteinView/components/IsoformSequencesToggle.d.ts

@@ -1,10 +1,7 @@
 import React from 'react';
-import type { Feature } from '@jbrowse/core/util';
+import type { IsoformSequences } from '../utils/util';
 export default function IsoformSequencesToggle({ structureSequence, structureName, isoformSequences, }: {
     structureSequence: string;
     structureName: string;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
+    isoformSequences: IsoformSequences;
 }): React.JSX.Element;

package/dist/LaunchProteinView/components/MSATable.d.ts

@@ -1,10 +1,7 @@
 import React from 'react';
-import type { Feature } from '@jbrowse/core/util';
+import type { IsoformSequences } from '../utils/util';
 export default function MSATable({ structureName, structureSequence, isoformSequences, }: {
     structureName: string;
     structureSequence: string;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
+    isoformSequences: IsoformSequences;
 }): React.JSX.Element;

package/dist/LaunchProteinView/components/TabPanel.js

@@ -1,5 +1,7 @@
 import React from 'react';
-// this is from MUI example
+// Panels stay mounted and are hidden via the `hidden` attribute rather than
+// unmounted, so switching tabs preserves each tab's in-progress work (typed
+// UniProt ID, fetched results, selected transcript) instead of resetting it.
 export default function TabPanel({ children, value, index, ...other }) {
-    return (React.createElement("div", { role: "tabpanel", hidden: value !== index, ...other }, value === index && React.createElement("div", null, children)));
+    return (React.createElement("div", { role: "tabpanel", hidden: value !== index, ...other }, children));
 }

package/dist/LaunchProteinView/components/TranscriptSelector.d.ts

@@ -1,4 +1,5 @@
 import React from 'react';
+import type { IsoformSequences } from '../utils/util';
 import type { Feature } from '@jbrowse/core/util';
 export default function TranscriptSelector({ val, setVal, isoforms, isoformSequences, structureSequence, feature, disabled, }: {
     isoforms: Feature[];
@@ -6,9 +7,6 @@ export default function TranscriptSelector({ val, setVal, isoforms, isoformSeque
     val: string | undefined;
     setVal: (str: string) => void;
     structureSequence?: string;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
+    isoformSequences: IsoformSequences;
     disabled?: boolean;
 }): React.JSX.Element;

package/dist/LaunchProteinView/components/TranscriptSelector.js

@@ -1,42 +1,26 @@
 import React from 'react';
 import { MenuItem, TextField } from '@mui/material';
-import { getGeneDisplayName, getTranscriptDisplayName, stripStopCodon, } from '../utils/util';
+import { classifyIsoforms, getGeneDisplayName, getTranscriptDisplayName, } from '../utils/util';
 export default function TranscriptSelector({ val, setVal, isoforms, isoformSequences, structureSequence, feature, disabled, }) {
     const geneName = getGeneDisplayName(feature);
-    const matches = [];
-    const nonMatches = [];
-    const noData = [];
-    for (const f of isoforms) {
-        const entry = isoformSequences[f.id()];
-        if (!entry) {
-            noData.push(f);
-        }
-        else if (structureSequence &&
-            stripStopCodon(entry.seq) === structureSequence) {
-            matches.push(f);
-        }
-        else {
-            nonMatches.push(f);
-        }
-    }
-    const byLengthDesc = (a, b) => isoformSequences[b.id()].seq.length - isoformSequences[a.id()].seq.length;
+    const { matches, nonMatches, noData } = classifyIsoforms({
+        options: isoforms,
+        isoformSequences,
+        structureSequence,
+    });
+    const renderOption = ({ feature: f, length }, note = '') => (React.createElement(MenuItem, { value: f.id(), key: f.id() },
+        geneName,
+        " - ",
+        getTranscriptDisplayName(f),
+        " (",
+        length,
+        "aa)",
+        note));
     return (React.createElement(TextField, { value: val ?? '', onChange: event => {
             setVal(event.target.value);
         }, label: "Choose transcript isoform", select: true, disabled: disabled },
-        matches.toSorted(byLengthDesc).map(f => (React.createElement(MenuItem, { value: f.id(), key: f.id() },
-            geneName,
-            " - ",
-            getTranscriptDisplayName(f),
-            " (",
-            isoformSequences[f.id()].seq.length,
-            "aa) (matches structure residues)"))),
-        nonMatches.toSorted(byLengthDesc).map(f => (React.createElement(MenuItem, { value: f.id(), key: f.id() },
-            geneName,
-            " - ",
-            getTranscriptDisplayName(f),
-            " (",
-            isoformSequences[f.id()].seq.length,
-            "aa)"))),
+        matches.map(m => renderOption(m, ' (matches structure residues)')),
+        nonMatches.map(m => renderOption(m)),
         noData.map(f => (React.createElement(MenuItem, { value: f.id(), key: f.id(), disabled: true },
             geneName,
             " - ",

package/dist/LaunchProteinView/hooks/useAlphaFoldDBSearch.d.ts

@@ -20,14 +20,8 @@ export default function useAlphaFoldDBSearch({ feature, view, }: {
     setUserSelection: import("react").Dispatch<import("react").SetStateAction<string | undefined>>;
     transcriptOptions: Feature[];
     selectedTranscript: Feature | undefined;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }> | undefined;
-    userSelectedProteinSequence: {
-        feature: Feature;
-        seq: string;
-    } | undefined;
+    isoformSequences: import("../utils/util").IsoformSequences | undefined;
+    userSelectedProteinSequence: import("../utils/util").IsoformSequence | undefined;
     uniprotEntries: import("../services/lookupMethods").UniProtEntry[];
     recognizedIds: string[];
     geneName: string | undefined;

package/dist/LaunchProteinView/hooks/useIsoformProteinSequences.d.ts

@@ -1,3 +1,4 @@
+import type { IsoformSequences } from '../utils/util';
 import type { Feature } from '@jbrowse/core/util';
 export default function useIsoformProteinSequences({ feature, view, }: {
     feature: Feature;
@@ -6,9 +7,6 @@ export default function useIsoformProteinSequences({ feature, view, }: {
     };
 }): {
     isLoading: boolean;
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }> | undefined;
+    isoformSequences: IsoformSequences | undefined;
     error: any;
 };

package/dist/LaunchProteinView/hooks/useTranscriptIsoformSelection.d.ts

@@ -8,17 +8,11 @@ export default function useTranscriptIsoformSelection({ feature, view, structure
     resetKey?: string;
 }): {
     transcripts: Feature[];
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }> | undefined;
+    isoformSequences: import("../utils/util").IsoformSequences | undefined;
     isLoading: boolean;
     error: any;
     selectedTranscriptId: string | undefined;
     setSelectedTranscriptId: import("react").Dispatch<import("react").SetStateAction<string | undefined>>;
     selectedTranscript: Feature | undefined;
-    selectedIsoform: {
-        feature: Feature;
-        seq: string;
-    } | undefined;
+    selectedIsoform: import("../utils/util").IsoformSequence | undefined;
 };

package/dist/LaunchProteinView/hooks/useTranscriptSelection.d.ts

@@ -1,10 +1,8 @@
+import type { IsoformSequences } from '../utils/util';
 import type { Feature } from '@jbrowse/core/util';
 export default function useTranscriptSelection({ options, isoformSequences, structureSequence, resetKey, }: {
     options: Feature[];
-    isoformSequences?: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
+    isoformSequences?: IsoformSequences;
     structureSequence?: string;
     resetKey?: string;
 }): {

package/dist/LaunchProteinView/hooks/useTranscriptSelection.js

@@ -1,4 +1,4 @@
-import { useMemo, useState } from 'react';
+import { useState } from 'react';
 import { selectBestTranscript } from '../utils/util';
 export default function useTranscriptSelection({ options, isoformSequences, structureSequence, resetKey, }) {
     const [userSelection, setUserSelection] = useState();
@@ -7,12 +7,12 @@ export default function useTranscriptSelection({ options, isoformSequences, stru
         setPrevResetKey(resetKey);
         setUserSelection(undefined);
     }
-    const autoSelection = useMemo(() => isoformSequences !== undefined
+    const autoSelection = isoformSequences !== undefined
         ? selectBestTranscript({
             options,
             isoformSequences,
             structureSequence,
         })?.id()
-        : undefined, [options, structureSequence, isoformSequences]);
+        : undefined;
     return { userSelection: userSelection ?? autoSelection, setUserSelection };
 }

package/dist/LaunchProteinView/services/lookupMethods.js

@@ -1,5 +1,5 @@
 import { jsonfetch } from '../../fetchUtils';
-import { getDatabaseTypeForId, isRecognizedDatabaseId, stripTrailingVersion, } from '../utils/util';
+import { buildUniProtXrefQuery, isRecognizedDatabaseId, stripTrailingVersion, } from '../utils/util';
 const UNIPROT_FIELDS = 'accession,id,gene_names,organism_name,protein_name,reviewed';
 function mapApiResultToEntry(result) {
     return {
@@ -11,31 +11,13 @@ function mapApiResultToEntry(result) {
         isReviewed: result.entryType === 'UniProtKB reviewed (Swiss-Prot)',
     };
 }
-/**
- * Build UniProt xref query for a recognized database ID
- */
-function buildXrefQuery(id) {
-    const dbType = getDatabaseTypeForId(id);
-    switch (dbType) {
-        case 'ensembl':
-            return `xref:ensembl-${id}`;
-        case 'refseq':
-            return `xref:refseq-${id}`;
-        case 'ccds':
-            return `xref:ccds-${id}`;
-        case 'hgnc':
-            return `xref:hgnc-${id.replace('HGNC:', '')}`;
-        default:
-            return undefined;
-    }
-}
 async function searchUniProt(query, size = 10) {
     const url = `https://rest.uniprot.org/uniprotkb/search?query=${encodeURIComponent(query)}&fields=${UNIPROT_FIELDS}&size=${size}`;
     const data = await jsonfetch(url);
     return data.results.map(mapApiResultToEntry);
 }
 async function searchByXref(id) {
-    const query = buildXrefQuery(id);
+    const query = buildUniProtXrefQuery(id);
     if (!query) {
         return { entries: [], error: undefined };
     }

package/dist/LaunchProteinView/utils/launchViewUtils.js

@@ -94,11 +94,13 @@ export async function launch1DProteinView({ session, view, feature, selectedTran
     });
 }
 // CROSS-REPO DEPENDENCY: the 'MsaView' view type is registered by
-// jbrowse-plugin-msaview, which wraps the `react-msaview` library. The `init`
-// keys below (msaUrl, colorSchemeName) and the connected* props are a runtime
-// contract with that plugin's model — they are NOT type-checked here because we
-// only depend on it at runtime (gated by hasMsaViewPlugin()). If react-msaview
-// renames these, the launch silently degrades. Keep in step with that repo.
+// jbrowse-plugin-msaview, which wraps the `react-msaview` library. The top-level
+// props here (colorSchemeName, connectedViewId, connectedFeature) are native
+// react-msaview model properties applied directly from the snapshot; only `init`
+// (msaUrl) is a declarative launch contract that the plugin resolves once and
+// clears. These are NOT type-checked here because we only depend on it at runtime
+// (gated by hasMsaViewPlugin()). If react-msaview renames these, the launch
+// silently degrades. Keep in step with that repo.
 export function launchMsaView({ session, view, feature, selectedTranscript, uniprotId, displayName, }) {
     if (!uniprotId) {
         return undefined;
@@ -109,9 +111,9 @@ export function launchMsaView({ session, view, feature, selectedTranscript, unip
             formatViewName('MSA view', feature, selectedTranscript, uniprotId),
         connectedViewId: view.id,
         connectedFeature: selectedTranscript?.toJSON(),
+        colorSchemeName: 'percent_identity',
         init: {
             msaUrl: getAlphaFoldMsaUrl(uniprotId),
-            colorSchemeName: 'percent_identity',
         },
     });
 }

package/dist/LaunchProteinView/utils/util.d.ts

@@ -6,14 +6,9 @@ export declare function getId(val?: Feature): string;
 export declare function getTranscriptDisplayName(val?: Feature): string;
 export declare function getGeneDisplayName(val?: Feature): string;
 export declare function getUniProtIdFromFeature(f?: Feature): string | undefined;
-/**
- * Check if an ID is a recognized database identifier that UniProt can map
- */
 export declare function isRecognizedDatabaseId(id: string): boolean;
-/**
- * Get the database type for a recognized ID (used for UniProt xref queries)
- */
-export declare function getDatabaseTypeForId(id: string): string | undefined;
+export declare function getDbIdLabel(id: string): string;
+export declare function buildUniProtXrefQuery(id: string): string | undefined;
 export declare function findRecognizedDbIds(f?: Feature): string[];
 export interface FeatureIdentifiers {
     recognizedIds: string[];
@@ -28,11 +23,27 @@ export interface FeatureIdentifiers {
  * geneId and geneName are always extracted from the parent feature 'f'.
  */
 export declare function extractFeatureIdentifiers(f?: Feature): FeatureIdentifiers;
-export declare function selectBestTranscript({ options, isoformSequences, structureSequence, }: {
+export interface IsoformSequence {
+    feature: Feature;
+    seq: string;
+}
+export type IsoformSequences = Record<string, IsoformSequence>;
+export interface RankedIsoform {
+    feature: Feature;
+    length: number;
+}
+export interface ClassifiedIsoforms {
+    matches: RankedIsoform[];
+    nonMatches: RankedIsoform[];
+    noData: Feature[];
+}
+export declare function classifyIsoforms({ options, isoformSequences, structureSequence, }: {
+    options: Feature[];
+    isoformSequences: IsoformSequences;
+    structureSequence?: string;
+}): ClassifiedIsoforms;
+export declare function selectBestTranscript(args: {
     options: Feature[];
-    isoformSequences: Record<string, {
-        feature: Feature;
-        seq: string;
-    }>;
-    structureSequence: string | undefined;
+    isoformSequences: IsoformSequences;
+    structureSequence?: string;
 }): Feature | undefined;

package/dist/LaunchProteinView/utils/util.js

@@ -34,48 +34,36 @@ export function getUniProtIdFromFeature(f) {
     }
     return f.get('uniprot') ?? f.get('uniprotId') ?? f.get('uniprotid');
 }
-// Ensembl ID patterns - covers human (ENS), mouse (ENSMUS), zebrafish (ENSDAR), etc.
-const ensemblGenePattern = /^ENS[A-Z]*G\d+/i;
-const ensemblTranscriptPattern = /^ENS[A-Z]*T\d+/i;
-const ensemblProteinPattern = /^ENS[A-Z]*P\d+/i;
-// NCBI RefSeq ID patterns
-const refSeqTranscriptPattern = /^[NX][MR]_\d+/i;
-const refSeqProteinPattern = /^[NX]P_\d+/i;
-// CCDS pattern
-const ccdsPattern = /^CCDS\d+/i;
-// HGNC pattern (HGNC:12345)
-const hgncPattern = /^HGNC:\d+/i;
-/**
- * Check if an ID is a recognized database identifier that UniProt can map
- */
+const DB_ID_PATTERNS = [
+    { db: 'ensembl', pattern: /^ENS[A-Z]*G\d+/i, label: 'Ensembl gene' },
+    { db: 'ensembl', pattern: /^ENS[A-Z]*T\d+/i, label: 'Ensembl transcript' },
+    { db: 'ensembl', pattern: /^ENS[A-Z]*P\d+/i, label: 'Ensembl protein' },
+    { db: 'refseq', pattern: /^[NX]M_\d+/i, label: 'RefSeq mRNA' },
+    { db: 'refseq', pattern: /^[NX]R_\d+/i, label: 'RefSeq ncRNA' },
+    { db: 'refseq', pattern: /^[NX]P_\d+/i, label: 'RefSeq protein' },
+    { db: 'ccds', pattern: /^CCDS\d+/i, label: 'CCDS' },
+    { db: 'hgnc', pattern: /^HGNC:\d+/i, label: 'HGNC' },
+];
+function matchDbIdPattern(id) {
+    return DB_ID_PATTERNS.find(p => p.pattern.test(id));
+}
+// Check if an ID is a recognized database identifier that UniProt can map
 export function isRecognizedDatabaseId(id) {
-    return (ensemblGenePattern.test(id) ||
-        ensemblTranscriptPattern.test(id) ||
-        ensemblProteinPattern.test(id) ||
-        refSeqTranscriptPattern.test(id) ||
-        refSeqProteinPattern.test(id) ||
-        ccdsPattern.test(id) ||
-        hgncPattern.test(id));
+    return matchDbIdPattern(id) !== undefined;
 }
-/**
- * Get the database type for a recognized ID (used for UniProt xref queries)
- */
-export function getDatabaseTypeForId(id) {
-    if (ensemblGenePattern.test(id) ||
-        ensemblTranscriptPattern.test(id) ||
-        ensemblProteinPattern.test(id)) {
-        return 'ensembl';
-    }
-    if (refSeqTranscriptPattern.test(id) || refSeqProteinPattern.test(id)) {
-        return 'refseq';
-    }
-    if (ccdsPattern.test(id)) {
-        return 'ccds';
-    }
-    if (hgncPattern.test(id)) {
-        return 'hgnc';
-    }
-    return undefined;
+// Human-readable label for an ID, e.g. "ENST00000123 (Ensembl transcript)".
+// Unrecognized IDs are returned unadorned.
+export function getDbIdLabel(id) {
+    const match = matchDbIdPattern(id);
+    return match ? `${id} (${match.label})` : id;
+}
+// Build the UniProt xref query fragment for a recognized ID, e.g.
+// "xref:ensembl-ENST00000123". HGNC strips its redundant "HGNC:" prefix.
+export function buildUniProtXrefQuery(id) {
+    const match = matchDbIdPattern(id);
+    return match
+        ? `xref:${match.db}-${match.db === 'hgnc' ? id.replace('HGNC:', '') : id}`
+        : undefined;
 }
 /**
  * Parse dbxref attribute which can have formats like:
@@ -154,7 +142,7 @@ export function findRecognizedDbIds(f) {
         if (/^\d+$/.test(hgncStr)) {
             recognizedIds.push(`HGNC:${hgncStr}`);
         }
-        else if (hgncPattern.test(hgncStr)) {
+        else if (matchDbIdPattern(hgncStr)?.db === 'hgnc') {
             recognizedIds.push(hgncStr);
         }
     }
@@ -206,12 +194,35 @@ export function extractFeatureIdentifiers(f) {
         geneName: typeof geneName === 'string' ? geneName : undefined,
     };
 }
-export function selectBestTranscript({ options, isoformSequences, structureSequence, }) {
-    const exactMatch = options.find(f => structureSequence &&
-        stripStopCodon(isoformSequences[f.id()]?.seq ?? '') === structureSequence);
-    const longestWithData = options
-        .filter(f => !!isoformSequences[f.id()])
-        .toSorted((a, b) => isoformSequences[b.id()].seq.length -
-        isoformSequences[a.id()].seq.length)[0];
-    return exactMatch ?? longestWithData;
+// The single rule for ranking transcript isoforms against a structure, shared
+// by the picker UI and the auto-selection: partition by whether the translated
+// protein matches the structure residues, with each group ordered longest-first.
+export function classifyIsoforms({ options, isoformSequences, structureSequence, }) {
+    const matches = [];
+    const nonMatches = [];
+    const noData = [];
+    for (const feature of options) {
+        const entry = isoformSequences[feature.id()];
+        const ranked = { feature, length: entry?.seq.length ?? 0 };
+        if (!entry) {
+            noData.push(feature);
+        }
+        else if (structureSequence &&
+            stripStopCodon(entry.seq) === structureSequence) {
+            matches.push(ranked);
+        }
+        else {
+            nonMatches.push(ranked);
+        }
+    }
+    const byLengthDesc = (a, b) => b.length - a.length;
+    return {
+        matches: matches.toSorted(byLengthDesc),
+        nonMatches: nonMatches.toSorted(byLengthDesc),
+        noData,
+    };
+}
+export function selectBestTranscript(args) {
+    const { matches, nonMatches } = classifyIsoforms(args);
+    return (matches[0] ?? nonMatches[0])?.feature;
 }