Heterobranchia Tree of Life Prospects

Public sequence data can already carry a nuclear backbone for — of the — order-level groups of Heterobranchia and for — named species. A whole-tree project fails only at the root: the — groups with no nuclear data are lower heterobranchs, Mesoneura and Ringiculimorpha, and closing that gap is a sequencing job rather than a mining job.

Built from Grant Wass's inventory of every NCBI record under Heterobranchia (harvested 24 October 2026; taxonomy from WoRMS/MolluscaBase; accessions verified at NCBI, ENA and DDBJ). Counts are formally named species unless stated; sp., cf. and aff. records are excluded from species counts.

Named species with any sequence data

—

— verified accessions

Species with nuclear genome-scale data

—

— of species with data

Families with nuclear data

— of —

— of WoRMS-accepted families

Order-level groups with nuclear data

— of —

Missing: —

Species with a chromosome-level genome

—

— with an assembly of any level

Species with a COI barcode

—

— of them also have nuclear data (placement anchors)

Which major groups are covered

Backbone data by major clade

Topology follows Brenzinger, Schrödl and Kano (2021); lower heterobranchs are a paraphyletic grade. Each tip shows one representative animal. Each bar splits the clade's WoRMS-accepted families by the best data any of its records provide. Hover a row for the groups it contains and what they are.

Nuclear genome-scale data (genome, transcriptome, UCE or exon capture) Mitogenome only Mitochondrial fragments or RAD-seq only No sequence data

Order-level groups: named species per data type

Rows in tree order, top of the tree first, each with one representative animal and a line on what the group is. Count cells share one log colour scale. The families column shows how many of a group's WoRMS families have nuclear data, with the bar as the share.

Group	Representative	What they are	Clade	WoRMS families	Families with nuclear data	Species with data	Nuclear	Genome	Chromosome-level	Transcriptome	UCE	Exon capture	Mitogenome	COI	RAD-seq taxa

Where the gaps are and what kind they are

Every WoRMS-accepted family, by its best data

One square per family, grouped by major clade and ordered by data layer, then by order-level group. Hover a square for the family.

Nuclear genome-scale data Mitogenome only Mitochondrial fragments or RAD-seq only No sequence data

The nuclear layer is four datasets that barely overlap

Named species with nuclear genome-scale data, by the combination of source types they have. Species with a single source type are highlighted.

One source type Two or more source types

Species by data layer

How the layers of a layered design overlap at species level.

Layer	Named species

Rooting gap

— of — groups have no nuclear data

They are —: minute, rarely collected marine snails that sit at or near the root of the tree. Between them they hold — WoRMS families and only — named species with any sequence, — of them by COI.

Nature: nothing to mine. Closing it means UCE capture from museum material. Without it a whole-Heterobranchia tree is rooted on distant caenogastropod outgroups.

Barcode-rich, backbone-poor giant

— of Stylommatophora species with data have nuclear data

Land snails are — of the named species in the inventory (—), but only — have genome-scale data. — of their — families are known only from mitochondrial sequence and — have none.

Nature: a placement exercise on a thin scaffold. A Stylommatophora backbone is a project of its own; in a Heterobranchia-wide tree most land-snail families would hang by mitochondrial signal.

Fragmented nuclear layer

— of nuclear-data species have one source type only

— of — species. Only — pair UCE capture with a genome or transcriptome, and — pair exon capture with one.

Nature: incompatible loci, not missing taxa. UCE, exon, transcriptome and genome sets become one matrix only after a common locus set is harvested in silico from every assembly.

Mitochondrial-only families

— families have mitochondrial data only

— have at least one mitogenome; — have only COI, 16S or other fragments, or RAD-seq reads. Together that is — of all accepted families.

Nature: placeable, not resolvable. These families can be hung on a nuclear backbone by placement, but their position rests on mitochondrial signal alone.

Unsampled families

— families have no sequence at all

— are land-snail families, — are nudibranchs and — are pteropods. WoRMS does not flag fossil-only families, so some of these are extinct; the rest are collection targets.

Nature: true absence. Only new specimens close these.

Uneven genome coverage

— of — groups have no genome assembly

— have no chromosome-level genome. The — chromosome-level genomes concentrate in Nudibranchia (—) and Stylommatophora (—); Cephalaspidea has —.

Nature: uneven rather than absent. Genomes are what probes are designed and validated against and what in-silico locus harvesting draws on, so these are the clades where new capture data would be hardest to design.

Whole tree or a subclade

How much of each candidate scope can sit on a nuclear backbone

Candidate scopes nested by clade. Each bar splits the scope's WoRMS-accepted families by best data layer; the figure at the end is the family count. Hover for counts.

Nuclear genome-scale data Mitogenome only Mitochondrial fragments or RAD-seq only No sequence data

Scope comparison

Named species counts; shares are of WoRMS-accepted families or genera in the scope.

Scope	Groups	With nuclear data	WoRMS families	Families with nuclear data	Families with any data	Genera with any data	Species with data	Nuclear	Chromosome-level genomes	COI	Anchors (nuclear and COI)	Groups lacking nuclear data

Reading the options

Whole Heterobranchia. A backbone is possible now for — of — groups and — of — families; — more families can be placed by mitochondrial data and — are unsampled. The payoff is a rooted, order-complete tree. The cost is a museum-sampling campaign across the — rootward groups before any deep node can be tested with heterobranch data.
Euthyneura. The clade PROJECT_PLAN targets. — of — groups have nuclear data (only —, a single family, lacks it), as do — of — families. — species have data, — with COI and — anchors. This is feasible from public data today, with the root taken from the literature and Acteonimorpha as the earliest branch.
A subclade now. Nudipleura has nuclear data for — of — families and — species, with — chromosome-level genomes and — COI species to place. Sacoglossa has — of — families, — nuclear species and — chromosome-level genomes. Either can be built from public data, and both are colour-relevant clades. Stylommatophora is the opposite case: — of — families and — of — species.

Definitions and what is left out

Named species: Formal binomials only. Records identified as sp., cf. or aff. are in the accession totals but not in species counts.
Nuclear genome-scale data: A genome assembly of any level, a transcriptome (TSA assembly or SRA RNA-Seq run), UCE target-capture reads, or exon and other nuclear target-capture reads.
Best data layer of a family: Nuclear genome-scale data if any record of the family has it; otherwise a mitogenome of 10 kb or more; otherwise mitochondrial gene fragments or RAD-seq reads; otherwise none. Grant's summary counts 129 mitochondrial-only families because it sets aside 10 families that have RAD-seq reads; here those count by their mitochondrial data, since RAD-seq reads do not contribute backbone loci.
Order-level group: The WoRMS order, or the nearest named rank above the family where WoRMS assigns no order (a superorder such as Sacoglossa or a superfamily such as Valvatoidea). The NCBI-only label Acochlidiacea is folded into Acochlidiimorpha. Fourteen accessions with no order and seven under Eupulmonata alone are in the accession total but in no group.
Clades: Lower heterobranchs, Mesoneura, Euthyneura, Acteonimorpha, Ringipleura, Tectipleura and Eupulmonata from the WoRMS classification; Euopisthobranchia and Panpulmonata and the branching order from Brenzinger, Schrödl and Kano (2021). Tetratentaculata is Ringipleura plus Tectipleura.
Denominators: WoRMS-accepted families (302) and genera (3,419) include some fossil-only taxa that WoRMS does not flag as extinct, so shares of families and genera with data are slightly understated.
Anchors: Named species that have both nuclear genome-scale data and a COI fragment. They let COI-only species be placed on a nuclear backbone.
Not in this inventory: Nuclear Sanger markers (H3, 18S, 28S), which PROJECT_PLAN lists as layer L1 and which still need to be mined from GenBank.
Representatives: One animal per group chosen as a familiar example, not a type species; the one-line descriptions are informal. Photos are the English Wikipedia page images of the species or genus article, with these credits: —

Standalone copy of the Claude dashboard Heterobranchia Tree of Life Prospects, built 2026-11-06. Data: genomic_data_inventory/raw/long_accessions.csv and the WoRMS tables beside it (Grant Wass, NCBI harvest of 24 October 2026), aggregated by visualizations/build/build_datasets.py and build_more.py; this file is produced by build_standalone.py. See visualizations/HANDOFF.md.