trpm7 review notes

Zebrafish trpm7 (UniProt Q563W7, TrEMBL/Unreviewed, 1774 aa) encodes TRPM7, a divalent-cation (Mg2+/Ca2+)-permeable channel fused to a C-terminal alpha-kinase ("chanzyme"). It is the touchtone/nutria/sweetbread/touchdown mutant locus in zebrafish. Core molecular functions are divalent-cation channel activity and alpha-kinase activity; downstream phenotypes span pigmentation, skeleton, kidney, pancreas, heart rate and touch/behavior.

GOA remapping note

This gene's GOA record is on the TrEMBL accession Q563W7. The refreshed GOA retains ~25 ZFIN experimental (IDA/IMP) rows (PMIDs 38970357, 27628598, 9007256, 23878236, 20881241, 21183474, 21832193, 17290233, 15823540) that were restored into goa.tsv by design; these are legitimate and reviewed on their merits. The refresh also dropped or re-referenced several broad IEA rows (see below).

Re-review 2026-09-29

• Reconciled 11 stale IEA rows against the refreshed GOA. Deleted 8 rows whose term/evidence/reference combination is no longer in GOA (GO:0000166 nucleotide binding, GO:0005261 monoatomic cation channel activity, GO:0016301 kinase activity, GO:0016740 transferase activity, GO:0034220 monoatomic ion transmembrane transport, GO:0046872 metal ion binding, and the GO_REF:0000043 IEA duplicates of GO:0005262 and GO:0006816 which remain represented as IBA rows). The three terms still present in GOA under a new reference (GO:0005524 ATP binding, GO:0006811 monoatomic ion transport, GO:0070588 calcium ion transmembrane transport) are now reviewed on their current GOA rows.
• Resolved the 4 current-GOA PENDING IEA rows as ACCEPT: GO:0005524 (ATP binding by the alpha-kinase domain), GO:0006811 (monoatomic ion transport), GO:0008324 (monoatomic cation transmembrane transporter activity), GO:0070588 (calcium ion transmembrane transport). Grounded on PMID:27628598 and PMID:20881241.
• Applied the participation test to the developmental/physiological IMP phenotype rows. TRPM7's molecular work is divalent-cation conductance and kinase activity; the tissue phenotypes are downstream consequences, so the following were changed ACCEPT/UNDECIDED -> KEEP_AS_NON_CORE: GO:0001966 thigmotaxis, GO:0036269 swimming behavior, GO:0048066 developmental pigmentation, GO:0002027 regulation of heart rate, GO:0003014 renal system process, GO:0031017 exocrine pancreas development, GO:0046928 regulation of neurotransmitter secretion, GO:0006582 melanin metabolic process, GO:0001503 ossification, GO:0030318 melanocyte differentiation, GO:0001501 skeletal system development. Kept as core (ACCEPT): GO:0010960 magnesium ion homeostasis (IDA), GO:0030001 metal ion transport (IDA), GO:0055080 monoatomic cation homeostasis (IMP), GO:0005886 plasma membrane (IDA) - these are the direct ion-transport/localization outputs. Rescue evidence supports the downstream reading, e.g. PMID:21183474.
• No REMOVE of experimental rows (phenotypes are genuine, only non-core). Rebuilt core_functions to three entries (divalent-cation/Mg transporter, calcium channel, alpha-kinase) and dropped the developmental process terms from directly_involved_in so it reflects the assigned actions. Retained the NEW GO:0015095 magnesium ion transmembrane transporter activity (TRPM7 is the principal Mg2+ conductance; passes participation test).
• Replaced deep-research-only support with verbatim primary quotes and added reference_review (relevance/correctness/notes) to all nine PMIDs. Validation: zero errors, zero warnings.

Follow-up after PR review (2026-10-04)

The nucleus (GO:0005634, IEA GO_REF:0000044) REMOVE reason previously claimed that no evidence supports nuclear localization, which contradicted the UniProt record's own ARBA prediction (Nucleus {ECO:0000256|ARBA:ARBA00004123}). The reason now states accurately that the only support is a statistical ARBA prediction and that the nuclear M7CK kinase fragment reported in mammalian cells is not evidence for full-length channel localization. The spliced deep-research quote reused under four localization rows was replaced with the verbatim sentence [file:DANRE/trpm7/trpm7-deep-research.md "Trpm7 is an integral membrane protein that primarily localizes to the plasma membrane of cells"].

Quote cleanup (2026-10-04)

• Fixed 12 supporting_text entries attributed to trpm7-deep-research.md that were ellipsis-spliced or not contiguous in the file: 8 were replaced with the contiguous source sentence/clause (restoring **/* markdown and the curly apostrophe where the source has them).
• The other 4 (GO:0098655, GO:0030001 IEA and IDA, GO:0010960, GO:0007346 rows: 'Mouse TRPM7 (mTRPM7) has been shown to best permeate ...' and 'Based on this profile ... primary mechanism for cellular Mg2+ homeostasis ...') were verbatim text from PMID:27628598, not from the deep-research file, so their reference_id was re-pointed to PMID:27628598 (already in references). Nothing dropped.
• No action, reason or term changed; just validate DANRE trpm7 passes with zero errors.

Follow-up after PR review (2026-10-09)

• Dropped the templated NEW row for GO:0015095 (IEA, no reference, "identified from core_functions analysis"). The specific activity is now expressed through the repo mechanism for a too-broad parent: the ARBA IEA GO:0008324 monoatomic cation transmembrane transporter activity row is MODIFY -> GO:0015095 magnesium ion transmembrane transporter activity (QuickGO confirms GO:0015095 is_a-descends from GO:0008324), supported by PMID:27628598 and PMID:21183474; the earlier quote about regulation BY intracellular Mg2+ was dropped as not evidence of Mg2+ transport.
• Rewrote the GO:0005634 nucleus summary so it matches its reason (only support is the ARBA prediction; the mammalian nuclear M7CK fragment is not evidence for the full-length channel), and the GO:0005886 plasma membrane IEA row now cites the actual evidence (HA-tagged drTRPM7 assessed for plasma membrane expression by whole-cell patch clamp in PMID:27628598, plus the UniProt SUBCELLULAR LOCATION line) instead of "Duplicate of IBA". The GO:0004672 IEA MODIFY row's "Duplicate of IBA" summary/reason was rewritten the same way. Validation: zero errors, zero warnings.

Follow-up after third-round PR review (2026-10-09)

Removed ancestor/descendant redundancy in core_functions: monoatomic cation homeostasis (GO:0055080) dropped from the Mg2+-transporter core function because magnesium ion homeostasis (GO:0010960) is its descendant (QuickGO is_a ancestors checked); calcium ion transmembrane transport (GO:0070588) moved from the Mg2+-transporter core function to the calcium-channel core function, where it replaces its own ancestors calcium ion transport (GO:0006816) and monoatomic cation transmembrane transport (GO:0098655). The GO:0005886 IDA row's bare reason was rewritten around the patch-clamp evidence in PMID:27628598. No annotation actions changed.

The reasons of the three parent rows dropped from core_functions (GO:0006816, GO:0098655, GO:0055080) now say they are retained as true parents whose core claim is carried by the specific descendant, so the prose matches the core_functions tree.