# MUC2 discrepancy resolved by explicit transcript alignment

REFERENCE-SEQUENCE CONSEQUENCE, NOT MEASURED EDITING OR HARM.

**The nominated hg38 site maps to NM_002457.5 codon4524, not3551. Its conditional single-base consequence is p.Tyr4524His.** The original p.Tyr3551His preserved the genomic residue identity but used the wrong full-transcript numbering; the uncapped p.Met3551Val used the full CDS at that wrong inherited number and therefore substituted an unrelated codon. Neither old accession-qualified call is correct for this nominated genomic position.

## Executed mapping

`muc2_exact_alignment.py` fetched and retained the exact primary records: UCSC hg38 `ncbiRefSeqPsl`, its `ncbiRefSeq` genePred representation, and NCBI NM_002457.5 GenBank. Total saved downloads are under60kB. `MUC2_EXACT_ALIGNMENT_RESULTS.json` records request URLs, file SHA-256, coordinates and assertions. No global scan or canonical edits occurred.

The unique overlapping accession PSL record has68 alignment blocks and preserves both genomic and transcript offsets. One-based hg38 chr11:1,103,307–1,103,309 maps uniquely to zero-based transcript13,597–13,599. NCBI gives CDS29–15,421 inclusive. Subtracting the28-base5-prime untranslated region yields zero-based CDS13,569–13,571, exactly codon4524. The transcript codon is TAC, independently matching the three genomic bases retrieved earlier. With the nominated minus-strand ABE guide, the plus/coding strand changes T-to-C: TAC→CAC, Tyr→His.

The fetched CDS has15,393nt and MD5 `c4722204402ceb72ee1510bdecdac7d2`, **identical to the historical CDS provenance table's NM_002457.5 entry**. Thus this is not explained by fetching a different CDS sequence today.

## Why the historical calculations disagreed

Concatenating the genomic coding exon segments from the genePred record reproduces codon3551 exactly. Using the PSL query coordinates instead produces codon4524, a2,919nt accumulated coordinate difference at this site. The transcript alignment contains substantial query insertions (2,915nt over the full alignment), so transcript length cannot be reconstructed by genomic exon-length summation alone. The full-alignment insertion total and the local representation offset are different quantities and are not equated here.

At full-CDS codon3551, the sequence really is ATG. That is why the uncapped producer computed M3551V. Its error was transferring the genome-concatenation codon number into the accession CDS without an alignment-aware coordinate mapping, then accepting either A or T instead of enforcing strand-specific edit direction. The genomic annotation's4,157aa versus full-CDS5,130aa protein lengths were an informative warning, but the PSL mapping supplies the actual resolution.

## What is resolved and what remains

The MUC2 coordinate/sequence discrepancy is resolved for the retained accession and public hg38 alignment. The previous report's “full-transcript residue unresolved” status is superseded by this evidence. No replacement of historical files is authorized or performed; use this new result when interpreting the nominated site.

This does not prove MUC2 is edited by the candidate, how often editing occurs, or whether Tyr4524His is harmful. Patient-specific repeat structure is not inferred. Nor does it independently certify every other mapping:9,247 other calls pass the newly executed direction check, but that alone cannot exclude same-base coordinate errors. Complete precursor CDS-cache recovery and full fresh-directory reproduction remain separate provenance work.

Sources are the exact primary URLs and hashes retained in the accompanying JSON. UCSC track metadata dates the alignment to2025-08-13; NCBI identifies accession.version explicitly. All assertions completed successfully.