Skip to content

MIxS 6.0.0 to 7.0.0: schema diff summary

Compares mixs6.0.0 (2022-03-24, 74744ee) to main (2026-07-29, 60b48ef), whose version is 7.0.0. Source: schema_comparison_results.yaml in this folder.

Read the headline numbers with care

The tool counts 409 new slots, 60 new classes, 770 slot definition changes and 287 class definition changes. Almost all of that is one refactor, not new metadata content:

  • 344 of the 409 "new" slots (84%) are *_data container slots, one per checklist-by-extension combination class (migs_ba_soil_data, mimarks_c_misip_water_data, and so on). Each has domain: MixsCompliantData, range: <the combination class>, multivalued: true, inlined: true. They are the machinery for a new top-level MixsCompliantData holder class. Only 65 new slots are actual metadata terms.
  • All 287 class definition changes are renames. Every one has an entry in expected_mappings, and the only fields that differ are name, class_uri (287 classes gained one), title (287 gained one) and description (281).
  • All 125 enum definition changes are renames too (snake_case_enum to CamelCaseEnum). Only one has any other change: pred_genome_type_enum -> ViralGenomeTypeEnum also changed its description.
  • 628 of the 644 multivalued changes and 158 of the 553 range changes are the diff seeing a dropped-but-unchanged default. v7 sets default_range: string at the schema level and stops writing multivalued: false explicitly, so range: string -> (absent) and multivalued: false -> (absent) mean nothing changed.

What the numbers understate: the 26 confirmed slot renames plus at least two unmapped ones (below) would otherwise read as removals of terms that in fact survived.

Added

Classes (60)

Group Count What it is
MimsMisip* 24 New MIMS-MISIP checklist ("Metagenome or Environmental with SIP") plus its 23 extension combinations
MimarksCMisip* 24 New MIMARKS specimen-MISIP checklist plus its 23 extension combinations
Mims*Ancient 8 The new Ancient extension combined with MIMS host-associated, human-associated, human-gut, human-oral, human-skin, plant-associated, sediment, soil
Ancient 1 New extension for degraded/ancient nucleic acids (alias minas)
Checklist, Extension, MixsCompliantData 3 Structural classes with no v6 counterpart. Extension carries alias EnvironmentalPackage

So the substantive additions are two checklists (MIMS-MISIP, MIMARKS-specimen-MISIP) and one extension (Ancient); the other 56 classes are combinations and scaffolding.

Slots: 344 combination containers + 65 real terms

The 65 real terms fall into four coherent themes, each arriving with its own enums:

Ancient DNA and palaeogenomics (16 terms) — matches the new Ancient extension: biocultural_label, chrono_age_protocol, chrono_age_remarks, context_retrieval_date, cultural_era, damage_treatment, earliest_chrono_age, earliest_chrono_sys, geological_epoch, host_preserv_state, latest_chrono_age, latest_chrono_sys, palaeopath_status, past_env_broad, past_env_local, stratigraph_context. New enums: BioCulturalLabelEnum, ChronoAgeProtocolEnum, ChronoAgeSysEnum, DamageTreatmentEnum, GeolEpochEnum. New prefix chrono: (TDWG chronometric-age terms).

Stable isotope probing (14 terms) — matches the new MISIP checklists: gradient_position, gradient_pos_density, gradient_pos_rel_am, internal_standard, isotope, isotopolog, isotopolog_approach, isotopolog_atom_frac, isotopolog_atom_pos, isotopolog_dose, isotopolog_incu_time, isotopolog_label, nucleobase_atom_frac, sip_method. New enums: IsotopeEnum, IsotopologApproachEnum, IsotopologLabelEnum.

Library prep, capture enrichment and read processing (14 terms): capt_pcr_cyc_tot, capt_probe_desc, capt_probe_src_taxid, reamp_pcr_cyc_tot, lib_gener_technique, lib_mid_desc, lib_polymerase, lib_strandedness, library_name, sop_lib_preparation, reads_removed, data_preproc_desc, marker_gene_recov, marker_gene_recov_sw. New enums: LibStrandEnum, LibTypeEnum.

Sample provenance, permits and identifiers (18 terms): permit_authority, permit_date, permit_id, permit_scope, orig_site_lat, orig_site_lon, orig_site_loc, orig_site_name, batch_ids, samp_alt_lab_ids, samp_category, samp_decont_pretreat, samp_dna_conc, samp_preserv_treatm, nucl_acid_extr_date, nucleic_acid_elution_vol, prev_pubs, sop_experimental. New enum: SampCategoryEnum.

The remaining three: estimated_genome_size (see missed renames), plus nose_mouth_teeth_throat_disord and urobiom_sex.

Enums (27 new)

The four theme clusters above account for 12. The rest are enumerations replacing free text on existing slots (see "range changes"), plus InsdcMissingValueEnum (INSDC missing-value vocabulary) and consolidations such as MoldVisibilityEnum, reused by five mold slots.

Prefixes (7 new), settings (33 new), subsets (4 new)

New prefixes: NCIT, SO, chrono, dc, schema, shex, xsd. The 33 new settings (float, unit, termID, termLabel, software, version, duration, timestamp, ...) are interpolation variables for LinkML structured_pattern, which v7 uses 307 times. This is the mechanism behind the 296 added pattern values.

New subsets: environment, investigation, nucleic acid sequence source, sequencing.

Removed

Classes (2): core, quantity value. The removal of quantity value is the single largest driver in the whole diff (173 range changes, 178 added patterns) and is discussed below.

Slots (18), in four groups:

Group Slots Status
Grouping slots turned into subsets environment field, investigation field, nucleic acid sequence source field, sequencing field Recorded in assets/between_diff_mappings/6_to_pre_7/inter_type_refactoring.tsv as slot-to-subset conversions
Grouping slots with no recorded fate core field, mixs extension field Same shape as the four above but absent from inter_type_refactoring.tsv
Attributes of the removed quantity value class has numeric value, has raw value, has unit Gone with the class
Recorded deletions soil_text_measure, texture_meth, tot_n_meth, tot_phos, url Listed in slot_name_mappings.tsv with an empty target
Not recorded anywhere estimated_size, soil_depth, samp_salinity, salinity_meth See missed renames

The five "recorded deletions" are not really deletions: each was a v6 duplicate sharing a MIXS URI with another v6 slot, and v7 kept one of the pair. Verified from the v6 terms.yaml and v7 src/mixs/schema/mixs.yaml:

v6 duplicate pair Shared URI v7 survivor
soil_text_measure, texture MIXS:0000335 soil_texture
texture_meth, soil_texture_meth MIXS:0000336 soil_texture_meth
tot_n_meth, tot_nitro_cont_meth MIXS:0000338 tot_nitro_cont_meth
tot_phos, tot_phosphate MIXS:0000689 tot_phosphate
url, associated resource MIXS:0000091 associated_resource

In each case the mapping file records the surviving name as a rename target for one member of the pair and blanks the other. That is defensible, but a consumer reading the diff alone sees five term deletions where the identifier actually survived under a different label.

Enums (27 removed). 18 belong to slots that dropped their enumerated range entirely (see range changes); the other 9 (add_recov_method_enum, assembly_software_enum, compl_score_enum, door_loc_enum, food_quality_date_enum, pres_animal_insect_enum, samp_purpose_enum, wall_loc_enum, window_loc_enum) belong to slots that moved to pattern-constrained strings.

Subsets (2): checklist, package (package is now expressed as the Extension class, which carries the alias EnvironmentalPackage).

Renamed

Classes: 287 renames, all following one convention — space-separated lowercase names become CamelCase, with the checklist abbreviated: MIGS bacteria -> MigsBa, MIGS eukaryote -> MigsEu, MIGS org -> MigsOrg, MIGS plant -> MigsPl, MIGS virus -> MigsVi, MIMARKS specimen -> MimarksC, MIMARKS survey -> MimarksS, MIMS -> Mims, MIMAG -> Mimag, MISAG -> Misag, MIUVIG -> Miuvig. Combination classes concatenate the two: agriculture MIGS bacteria -> MigsBaAgriculture, with the checklist moving to the front.

Enums: 125 renames, all snake_case_enum to CamelCaseEnum. Six of the new names absorb more than one old enum: CeilingWallTextureEnum (2), SoilHorizonEnum (2), GeolAgeEnum (2), CompassDirections8Enum (2, from ext_wall_orient_enum + ext_window_orient_enum), DamagedEnum (3), DamagedRupturedEnum (3). So 125 old enums map onto 117 new ones.

Subsets: 1 — checklist_package_combination -> combination_classes.

Slots: 26 confirmed renames, driven by three consistent conventions rather than ad hoc edits:

  1. Abbreviate to the established MIxS style: assembly_quality -> assembly_qual, microbial_biomass_meth -> micro_biomass_meth, previous_land_use_meth -> prev_land_use_meth, non_mineral_nutr_regm -> non_min_nutr_regm, host_infra_specific_name -> host_infra_spec_name, host_infra_specific_rank -> host_infra_spec_rank, host_family_relation -> host_fam_rel, single_cell_lysis_appr -> sc_lysis_approach, single_cell_lysis_prot -> sc_lysis_method, air particulate matter concentration -> air_PM_concen, tot_car -> tot_carb.
  2. Normalize the samp_ family, expanding truncated stems and standardizing on samp_: samp_collec_device -> samp_collect_device, samp_collec_method -> samp_collect_method, sample_collec_method -> samp_collect_method, sample_name -> samp_name, samp_stor_dur -> samp_store_dur, samp_stor_loc -> samp_store_loc, samp_stor_temp -> samp_store_temp.
  3. Disambiguate generic names with a domain prefix: horizon -> soil_horizon, texture -> soil_texture.

Plus mechanical identifier fixes: x_16s_recover -> x16s_recover, x_16s_recover_software -> x16s_recover_software, Food_Product_type -> food_product_type, Food_source -> food_source, associated resource -> associated_resource.

Two old names collapse into one new one in each of two cases: host_family_relation and host_family_relationship both to host_fam_rel; samp_collec_method and sample_collec_method both to samp_collect_method.

One rename also moved its identifier: assembly_quality -> assembly_qual changed slot_uri from MIXS:0000058 to MIXS:0000056. That is the only slot_uri change on a renamed slot, and it deserves a maintainer's eye. Separately, prod_label_claims gained slot_uri: MIXS:0001337 where it previously had none.

Possible missed renames

This diff has no rename_candidates section, so the following are my own findings, each confirmed by MIXS URI identity between the v6 and v7 schema sources. A maintainer should confirm these and add the real ones to assets/between_diff_mappings/6_to_pre_7/.

v6 slot (reported as removed) Shared URI v7 slot (reported as added or shared) Evidence
estimated_size MIXS:0000024 estimated_genome_size Same URI. The class cascaded entries show estimated_size removed from exactly 144 classes and estimated_genome_size added to exactly 144. Not in slot_name_mappings.tsv at all.
soil_depth MIXS:0000018 depth Same URI. soil_depth was a v6 duplicate of depth; v7 kept depth. Not in slot_name_mappings.tsv at all, not even as a blank-target deletion like the other four duplicate cleanups.

estimated_size -> estimated_genome_size is the more consequential of the two: it is a real rename of a live term across 144 classes, currently presented as one deletion plus one addition.

Two further v6 slots have no v7 successor carrying their URI and appear to be genuine retirements, but they are absent from the mapping files entirely, so nothing records the decision: samp_salinity (MIXS:0000109) and salinity_meth (MIXS:0000341). Both were dropped from 24 classes each. salinity (MIXS:0000183) survives, so a reader may reasonably wonder whether these were meant to merge into it.

Finally, core field and mixs extension field look like the same slot-to-subset conversion recorded for the other four grouping slots, but no target subset exists for either. Worth a line in inter_type_refactoring.tsv even if the answer is "dropped".

Cardinality and range changes

The quantity value class is gone: 173 slots, the largest single change in the diff

v6 modeled measurements as a quantity value class with has numeric value, has raw value and has unit. v7 removes the class and replaces it with pattern-constrained strings. Concretely, 173 slots changed range: quantity value -> string and 178 slots that previously had range: quantity value gained the pattern ^[-+]?[0-9]*\.?[0-9]+(?:[eE][-+]?[0-9]+)?( *- *[-+]?[0-9]*\.?[0-9]+(?:[eE][-+]?[0-9]+)?)? *.*$ (a number, or a numeric range, optionally followed by a unit). This covers the chemistry, physical-measurement and morphometric terms: temp, depth, alt, elev, salinity, nitrate, phosphate, diss_oxygen, host_age, wind_speed, samp_size and so on.

This is a validation loosening, not a tightening. A structured object with a typed numeric field and a unit field became a free-text string checked by a regex that ends in .*. Anyone who parsed quantity value structurally must change their code, and the new form cannot enforce that the unit is meaningful.

A minority of former quantity value slots went to a real numeric type instead: rel_air_humidity, carb_nitro_ratio, surf_humidity, season_humidity, iwf, animal_am_freq (6 slots to float). ferm_ch-style percent slots and a few others moved string -> float: contam_score, ferm_headspace_oxy, ferm_chem_add_perc.

Other range changes worth listing individually

  • 127 enum-range renames (window_cond_enum -> DamagedRupturedEnum and similar). No semantic change; they follow the enum renames above.
  • 21 slots gained an enumerated range where they previously accepted any string. This is a genuine tightening: aero_struc, assembly_qual, built_struc_set, ceil_struc, ceil_water_mold, contam_screen_input, door_water_mold, fireplace_type, heat_sys_deliv_meth, host_dependence, season, seq_quality_check, shad_dev_water_mold, shading_device_loc, space_typ_state, sym_life_cycle_type, urine_collect_meth, wall_water_mold, wga_amp_appr, window_status, window_water_mold. Five of them (ceil_water_mold, door_water_mold, shad_dev_water_mold, wall_water_mold, window_water_mold) share one new MoldVisibilityEnum.
  • 18 slots lost their enumerated range entirely and now accept any string. This is a genuine loosening: cur_land_use, decontam_software, door_type_wood, drug_usage, floor_finish_mat, host_sex, microb_start_count, organism_count, plant_growth_med, room_type, samp_floor, samp_md, source_uvig, special_diet, spikein_count, study_complt_stat, vis_media, water_source_shared.
  • 17 date slots became datetime: collection_date, cult_isol_date, pregnancy, menarche, menopause, hrt, douche, flooding, fire, extreme_event, last_clean, date_last_rain, date_extr_weath, fertilizer_date, cons_purch_date, prod_start_date, iw_bt_date_well. Existing date-only values remain valid ISO 8601, so this widens what is accepted.
  • 9 double -> float (soil_pH, water_pH, root_med_ph, ferm_pH, surf_moisture_ph, ph, avg_occup, typ_occup_density, occup_density_samp) plus one integer -> float (occup_samp).
  • 8 integer -> string (a loosening on counts): number_plants, max_occup, number_pets, room_occup, exp_pipe, freq_cook, host_occupation, number_resident.
  • 6 string -> boolean (a tightening): reassembly_bin, hysterectomy, smoker, twin_sibling, medic_hist_perform, x16s_recover.
  • timepoint: double -> string, and samp_time_out: double -> (absent).
  • Ranges dropped without a replacement type, so they fall back to the new default_range: string: trnas (was integer), host_spec_range (was integer), host_of_host_taxid (was integer), ferm_chem_add (was quantity value). trnas and host_of_host_taxid losing integer looks like an oversight worth checking.

Multivalued

Of 644 flagged changes, 628 are false -> (absent) (no change). The 16 real ones:

  • Became multivalued (11): source_mat_id, experimental_factor, host_spec_range, ventilation_type, biotic_regm, env_medium, food_dis_point_city, sop, solar_irradiance, associated_resource (renamed), host_fam_rel (renamed). env_medium becoming multivalued is notable: it is a required core term that many archives treat as single-valued.
  • Stopped being multivalued (5): farm_water_source, food_clean_proc, photosynt_activ, food_product_type, soil_temp. These four-plus-one are a narrowing; any existing record with multiple values is no longer valid.

Required

36 slots gained required: true on the global slot definition. This is largely a relocation, not 36 new obligations. v6 never set required on a global slot (zero occurrences in the v6 terms.yaml); it set requiredness per class in slot_usage. For example env_broad_scale was required: true in 11 v6 checklist classes and is now declared required once on the slot.

The slots: abs_air_humidity, add_recov_method, api, basin, build_occup_type, building_setting, coll_site_geo_feat, collection_date, env_broad_scale, env_local_scale, env_medium, filter_type, geo_loc_name, hc_produced, hcr, heat_cool_type, host_dependence, IFSAC_category, indoor_space, iwf, lat_lon, light_type, occup_density_samp, occup_samp, project_name, rel_air_humidity, samp_collect_point, samp_name, samp_taxon_id, samp_type, seq_meth, space_typ_state, sym_life_cycle_type, typ_occup_density, water_cut, plus sample_name -> samp_name.

Where the diff reports a class-level requiredness change, it is confined to 23 classes and one removal, concentrated in two extensions: indoor_surf and surf_material became required in all 12 BuiltEnvironment classes, and samp_vol_we_dna_ext (11), samp_size (6), assembly_qual (5), lib_reads_seqd (5), lib_screen (5), lib_vector (5), samp_mat_process (5), host_spec_range (4), specific_host (2), pathogenicity (1) in Agriculture classes. isol_growth_condt stopped being required in MimarksSAgriculture. See the caveat in Notes before acting on these.

Pattern changes

301 slots changed a pattern: 296 added, 4 removed, 1 modified. The additions are not hand-written; they are materialized from 307 structured_pattern declarations that use the 33 new settings as interpolation variables. Only 27 distinct regexes cover all 296 slots:

Slots What it constrains
178 Number or numeric range, optional trailing unit (the quantity value replacement)
32 PMID:, doi:, or an http(s) URL (two near-identical variants; the citation/reference terms)
19 term label [PREFIX:id] (the ontology-term style used by env_broad_scale and similar)
13 Positive integer plus free text
9 ISO 8601 duration
7 text;number unit
6 Three semicolon-separated fields
5 Either an ontology term or a positive integer
4 each Free text or ontology term; text;positive integer
3 Numeric range plus unit
the rest One-off shapes such as 95% ANI; 85% AF; ... for otu_class_appr

Four slots lost their pattern, which is a loosening on already-structured fields: food_quality_date (was ^(best by|best if used by|freeze by||use by);YYYY-MM-DD$), contam_screen_param (was ^(ref db|kmer|coverage|combination);.+), compl_score (was ^(high|med|low);(0|[0-9]{1,2}|100)%$), pres_animal_insect (was ^(cat|dog|rodent|snake|other);\d+$). All four also lost their enum in the same release, so these terms went from doubly constrained to free text in one step. That may be intentional, but it is the sharpest validation regression in the diff and is easy to miss among 296 additions.

One pattern modified: add_recov_method. The old regex embedded a sprawling inline ISO 8601 datetime; the new one is a readable equivalent. The accepted-value list is unchanged, but the two regexes are not exactly equivalent at the edges (the old one accepted week dates like W05 and ordinal dates; the new one does not).

Title changes

ENA uses the MIxS title as its controlled nomenclature, so these are identifier changes for downstream consumers even where the edit looks like tidying. Eleven titles changed:

Meaning changed:

Slot Old title New title
nose_throat_disord lung/nose-throat disorder nose throat disorder
horizon_meth soil horizon method horizon method
sieving composite design/sieving sieving
horizon -> soil_horizon horizon soil horizon
texture -> soil_texture texture soil texture
air particulate matter concentration -> air_PM_concen air_PM_concen air particulate matter concentration

Note that horizon_meth moves the other way from horizon -> soil_horizon: the slot renamed toward soil_, its method partner renamed away from it. Worth checking that this is deliberate.

Case, spacing or spelling only (still visible to ENA users):

Slot Old title New title
food_trav_vehic Food shipping transportation vehicle food shipping transportation vehicle
samp_taxon_id Taxonomy ID of DNA sample taxonomy ID of DNA sample
Food_Product_type -> food_product_type Foodon product type food product type
Food_source -> food_source Food source food source
samp_transport_cont sample transport container (double space) sample transport container
host_of_host_env_med host of the symbiotic host environemental medium host of the symbiotic host environmental medium

Cosmetic changes (grouped)

293 slot descriptions changed. Only 60 carry any information:

  • 233: trailing period removed. A single mass edit stripping the final . from descriptions. Example: heavy_metals, ... add multiple copies of this field. to ... add multiple copies of this field.
  • 3: whitespace or case only. wind_speed ("Speed" to "speed"), ph_meth ("ph" to "pH"), ph (whitespace).
  • 13 of the 57 remaining changes are mojibake repair, not editorial: v6 had UTF-8 read as Latin-1 (¬∞ for °, ¬† for a non-breaking space) in api, aromatics_pc, asphaltenes_pc, pour_point, resins_pc, root_med_macronutr, root_med_micronutr, root_med_regl, root_med_suppl, samp_transport_cond, saturates_pc, tan, viscosity. All 13 are clean in v7. The repair replaced the corrupted degree signs with spaces rather than °, so api now reads e.g. 31.1 API.
  • 44 substantive rewrites. Two drivers dominate. First, de-soiling: definitions written for the soil package were generalized, e.g. tot_org_carb from "Definition for soil: total organic carbon content of the soil, definition otherwise: total organic carbon content" to "Total organic carbon content"; store_cond from "how and for how long the soil sample was stored" to "how and for how long the sample was stored", with added guidance. Second, typo and reference fixes: hall_count (cooridors to corridors), samp_transport_cont (Conatiner to Container), plant_growth_med (EO to PECO), otu_class_appr (OTUS to OTUs), lat_lon (added "limited to 8 decimal points"). One rewrite reverses a stated meaning: size_frac_up went from "Materials smaller than the size threshold are excluded" to "Materials larger than the size threshold are excluded" — worth confirming which is correct.

281 class descriptions also changed, but 253 of them are one template swap for the combination classes: "Combinatorial checklist with environmental package " became "MIxS Data that comply with the checklist and the Extension". Buried among them are four real corrections that the template swap hides:

  • MIGS org: "Minimal Information about a Genome Sequence: org" to "... organelle"
  • MIGS plant: "... plant" to "... plasmid"
  • MIGS virus: was "... cultured bacteria/archaea" (a copy-paste error in v6), now "... virus"
  • MIMARKS specimen and MIMARKS survey: "Minimal Information about a Marker Specimen" to "... Marker Sequence"

Notes

The MIXS: prefix now expands to a different IRI. v6 had MIXS: https://w3id.org/mixs/terms/ with default_prefix: mixs.vocab; v7 has MIXS: https://w3id.org/mixs/ with default_prefix: MIXS. Every slot_uri and class_uri in the schema therefore resolves to a different absolute IRI than it did in v6 (MIXS:0000018 was https://w3id.org/mixs/terms/0000018, now https://w3id.org/mixs/0000018). The diff reports this as one line in prefixes.definition_changes and it is easy to overlook, but it affects every RDF consumer. The schema id also changed from http://w3id.org/mixs to https://w3id.org/mixs (scheme only). Confirm this is intended and that the old IRIs still resolve.

The class-level cascaded entries are not trustworthy as content changes. The tool reports required added: assembly_qual for agriculture MIGS bacteria -> MigsBaAgriculture, but in v6 that class declared no slots at all: it got them via is_a: agriculture plus mixins: [MIGS bacteria], and MIGS bacteria already declared assembly_qual as required. The ancestors added: Extension (276) and ancestors added: Checklist (264) counts confirm the tool tracks is_a ancestry. I could not determine from the diff whether it resolves mixins, and the evidence above suggests it does not, or does so inconsistently. That would explain why the same slots appear as both added (adapters 25 times) and removed (adapters 23 times) across classes. Treat every cascaded entry, including the class-level required and recommended lists above, as a pointer to check against the schema rather than as a finding. The one cascade I did verify independently is estimated_size / estimated_genome_size (144 removals, 144 additions, same MIXS URI).

recommended shifted almost everywhere. elev (261 classes), depth (260), alt (256) and temp (253) became recommended nearly across the board. Same caveat as above.

No rename_candidates section. This tool does not emit one, so the "Possible missed renames" section is my own URI-identity analysis rather than the tool's. A rename_candidates pass in diff-releases would have caught estimated_size and soil_depth automatically.

Version label. The new side is tagged main, not v7.0.0; the version scalar reads 7.0.0 and the commit is 60b48ef. The comparison was generated from assets/releases_for_diffing/main_60b48ef, which is not among the two main_* snapshots currently in that folder (main_07b5360, main_1591e32). URI checks in this summary were run against the repository's working src/mixs/schema/mixs.yaml on branch release-v7.0.0-local, not against 60b48ef directly.

Left unclassified. Both nose_throat_disord and the new nose_mouth_teeth_throat_disord exist in v7, and their content crossed over: nose_mouth_teeth_throat_disord carries essentially v6's nose_throat_disord description (the Human Disease Ontology guidance), while nose_throat_disord's description was replaced with the shorter "Report any history of nose, mouth, teeth and/or throat disorders". That looks like a rename that left the old slot in place rather than a deliberate split, but nothing in the mapping files records it and I could not tell from the diff alone.