How should Neocaridina be selectively bred?

Evidence review: August 10, 2026

Selective breeding means deciding which documented animals contribute to the next generation, then measuring what their offspring actually become. Removing pale animals can make a sales group look more consistent today. It does not reveal which parent produces consistent offspring, measure genetic diversity or prove that a line will improve.

Six claims that must stay separate

99.3 percent core-gene recovery is not a 99.3 percent complete genome

The foundational 2014 Neocaridina genome paper sequenced DNA from one commercially sourced red-patched adult. Its short-read assembly contained 3,346,358 contigs, an N50 of 400 base pairs and 1.284 gigabases of assembled sequence. The authors compared that result with an estimated approximately 3-gigabase genome and said non-coding regions were probably poorly recovered.¹²

The often-tempting 99.3 percent figure means recognizable hits for 455 of 458 selected core eukaryotic genes at the paper's search threshold. It is a gene-detection benchmark, not a measure showing that 99.3 percent of the nuclear genome was assembled. The study also recovered a 15,565-base-pair mitochondrial genome, but mitochondrial sequence from one animal does not establish a seller line's pedigree, purity, colour inheritance or population-wide diversity.

This resource made later gene discovery possible. It did not produce a consumer genetic test, validate a commercial colour name or show which offspring two named hobby lines will produce. Ask which animals, markers and validation population a genetic claim actually used before treating a genome reference as a breeding assay.

A model-organism recommendation is not a care standard

A 2015 symposium policy paper recommended N. denticulata as a decapod functional-genomics model. The authors compared it with two crayfish species and emphasized its small space requirement, commercial availability, transparent cuticle, cited life-history traits and preliminary genome.¹⁵ The paper reported no new animals, aquarium comparison or statistical analysis, and it did not describe a systematic literature search.

Its pH 6.5 to 8.0, temperature up to 30 C, preferred 22 to 25 C and pH 7.0 to 7.5, 15 to 16 day intermolt, 20 to 30 egg, approximately 30 day hatch and 4 to 6 month maturity statements summarize selected older sources. They do not come from one standardized trial of a documented aquarium population and are not universal care ranges or clocks.

The paper also called a complete genome and reliable gene-control methods the next steps. Proposed reporter genes, transgenesis, mutant screening and targeted expression were future research directions, not demonstrated hobby tools. Its broad infection-resistance sentence likewise does not establish immunity to all bacteria, pathogen-free stock or a treatment. Use this paper to understand why researchers chose the animal, then return to the underlying experiment for every care or health claim.

A hormone-pathway gene is not a breeding treatment

A 2015 study used that draft genome to catalogue candidate genes associated with ecdysteroid and sesquiterpenoid pathways. Reciprocal sequence searches, domains, Bayesian phylogenetic analysis and selected cloned partial transcripts supported the catalogue. The authors also displayed a semiquantitative JHAMT gel using three reported individual samples per life-stage or body-region group.¹⁴

Sequence similarity does not prove that every candidate performs the inferred function in Neocaridina. No hormone or protein concentration, live-animal pathway manipulation, cross, molt outcome, growth, survival, fertility or offspring endpoint was measured. The JHAMT comparison reported no reference-gene normalization, band densitometry or statistical model. It cannot identify a breeder, predict fecundity, validate a commercial line or justify hormone dosing.

Three altered survivors are not an inheritance chart

A 2022 study removed one-cell embryos from females four hours after spawning and microinjected CRISPR/Cas9 material targeting Nh-scarlet. Among 28 injected embryos that survived to the eye-phenotype screen, two had altered eye shape and one had an absent eye. Target-region sequencing found insertions, replacements or deletions in the three visibly altered survivors.¹³

The paper calls this 10 percent gene-editing efficiency, but that is three divided by 28 surviving injected embryos. It is not 10 percent of every embryo originally injected because the starting denominator was not reported. The screen selected abnormal eyes, and the paper does not show systematic genotyping of every phenotype-normal survivor. Some embryos also carried more than one edited sequence, which is consistent with mosaicism.

This is direct evidence that targeted embryo editing and an early eye phenotype were technically possible under the laboratory method. It is not evidence of adult survival, normal health, body-colour control, fertility, germline transmission or F1 inheritance. No stable hobby line was created. An edited eye in an early survivor is therefore not a commercial-colour recipe, trade-line identity test or inheritance chart.

What colour-genetics research establishes

A 2026 study using red, yellow and blue strains identified pigment-pathway genes and different expression patterns among strains. Carotenoid and pteridine results differed between the blue and red or yellow groups, and the abundance of the astaxanthin-binding protein CRCN was higher in the tested blue strain.¹ The paper called its study animal Neocaridina denticulata.

Those red, yellow and blue labels were not connected to Blue Dream, Sunkist or another named aquarium line. The source reported three biological samples per endpoint, but not how many shrimp were pooled into each sample, how many culture tanks were used or how strains were allocated among tanks. It also reported no voucher, diagnostic species confirmation, source ancestry, standardized colorimetry, carotenoid composition or individual pteridine compounds. Its fluorescence result was relative and had no reported tissue-mass normalization. These results identify pathways worth testing. They do not rank genes causally, predict grade, prove inheritance or prescribe a colour food.

A separate 2026 study, using the name N. denticulata sinensis, used gene expression, RNA interference and targeted genotyping to study NinaB-like. It reported increased red pigment after knockdown and two tightly linked non-synonymous variants associated with body colour in its sampled strains.² These are important molecular results. They are not a consumer genetic test, a pedigree assay, a dominance table or proof that the variants diagnose Blue Dream, Sunkist, Bloody Mary or another seller label.

A functional gene result is still not a commercial-line assay

Huo et al. studied NdBCO-like4 expression, an artificial RNA-interference exposure and one candidate SNP in laboratory red, yellow, blue and wild or transparent strains.¹⁰ The strongest functional comparison used five mothers. Fifteen sibling embryos per well received either NdBCO-like4 or EGFP dsRNA after removal and membrane-softening treatment. Lower red-channel brightness and pigment-particle distribution followed the target treatment under that 24-hour protocol.

This does not validate the SNP as the cause of a trade phenotype. The abstract calls G.1719G>A synonymous, while the results and discussion call it a missense D445N change. The stated 192 genotypes per strain do not match the displayed totals of 148, 95, 190 and 172, and the same comparison strain is called both wild and transparent. Tank replication and family contributions to the strain comparison are not reported. The source cannot identify a named commercial line, establish parentage, supply F1 or F2 proportions, or turn one candidate marker into a breeding decision.

Zhang et al. tested another carotenoid candidate, scarb1, with expression, split-clutch RNA interference and a synonymous SNP association.¹¹ Five maternal clutches supplied the functional comparison. The target exposure changed metanauplius chromatophore measurements, while a later pre-zoea attempt did not silence the gene.

The reported yellow population was entirely GA at G1593A and the reported red population entirely AA, but association is not inheritance. The design planned 384 genotypes per population and reports 345, without accounting for the 39 missing calls in each population. It did not cross populations, score F1 or F2, validate an independent seller line, test individual prediction or demonstrate that the synonymous site caused the phenotype. The study also does not state the statistical method used for most expression and RNA-interference comparisons. This marker cannot diagnose Sunkist, Yellow Neon or any other commercial line.

Expression differences can be associated with colour without telling a hobbyist the expected F1 or F2 proportions from two named lines. The reviewed molecular papers did not run the documented named-morph crosses needed to validate the popular online colour-family charts. See what is known about mixed-colour tanks and crossing records.

Five SSR markers are not a colour-purity certificate

A 2020 study developed five polymorphic SSR loci from supplier stocks labelled super red and chocolate shrimp, then compared genotype frequencies in pure-stock and reciprocal-cross offspring.¹⁶ The supplement reports eight selected females and eight selected males from each source strain and three families in each of four cross groups. This is useful marker development within that breeding material.

It is not a validated test for a hobby trade name. The paper states 50 offspring per group, but its reciprocal-hybrid frequencies occur in exact eighths. It does not provide individual genotypes, missing-call denominators, family assignments, scored colour phenotypes, Mendelian segregation, linkage or an independent test population. The advertised 4.31 percent single-marker and 0.0018 percent combined misrecognition figures are not derived in the article or supplements and were not measured on a held-out sample.

The transcriptome comparison also appears to use one pooled library for each colour stock, not replicated biological libraries. The five loci may help a properly designed follow-up study, but they cannot certify purity, pedigree, ancestry, genetic diversity, Blue Dream, Sunkist, Bloody Mary or another seller line. The paper's super red label does not establish equivalence to a current red trade name.

Three colour labels were also three tank histories

A 2022 microscopy study described changing chromatophore types, forms and distributions in laboratory strains labelled red, yellow and blue from embryo through early post-hatch development.⁸ It can describe pigment-cell morphology. It cannot isolate inheritance because one strain was cultured in one tank and the families contributing the selected eggs were not reported.

The study ran no crosses, pedigree analysis, F1 or F2 counts, commercial-line validation or adult-grade follow-up. Its developmental differences therefore do not prove genetic purity, a named trade identity or a universal age at which a juvenile can be graded.

Twenty-four genes do not make a twenty-four-trait chart

A 2025 study identified 24 candidate crustacyanin genes in one unpublished laboratory genome, compared their family relationships and examined tissue expression and antibody localization.⁹ The supplement contains 15 expression libraries, three for each of five tissues, but does not identify the animal or pool behind each library. The paper also leaves total animal count, sex, age, colour line, family and independent immunohistochemistry units unreported.

No named line comparison, cross, pedigree, colour phenotype, genotype association, F1 or F2 ratio, or selection response was measured. A candidate family found in one source genome cannot establish population-wide copy number or tell a keeper which allele a commercial shrimp carries. The similar proteins may also be recognized together by the study's polyclonal antibody, so its tissue signal does not assign a separate function to each gene.

What one generation of full-sibling mating changed

Tropea, Marciano and López Greco compared females paired with brothers against females paired with non-brothers from the experimental stock. All surviving females in both groups mated and spawned. Total egg number and measured fertilization success in the first brood were similar. In a separate second-brood comparison, the mean number of newly hatched juveniles was 24.67 per female in full-sibling clutches and 31.43 in non-sibling clutches.³ Because the egg and hatch measures came from different broods, they do not form one hatch-rate calculation. The paper interpreted the lower hatch count as likely greater egg loss or lower embryonic survival in the sibling treatment.

The study followed juvenile quality for 60 days under ordinary feeding and for 20 days in a food-restriction experiment. It detected a later juvenile survival cost under the ordinary protocol. In the food experiment, length gain through day 10 was lower in the sibling treatment when averaged across all three feeding groups; that difference was no longer evident during days 10 to 20, and survival did not differ by breeding treatment. Other growth measures did not show the same pattern.³ This is direct evidence that one known full-sibling generation can affect some fitness outcomes. It does not show that every small colony is inbred, establish a minimum founder count, predict colour loss or prove that a periodic commercial outcross fixes the problem.

Relatedness can affect more than hatching

In a later experiment, females housed with brothers took longer to mature ovaries than females housed with non-brothers. Eggs from females mated with brothers had lower carotenoid content, higher cholesterol and a tendency toward lower energy content. Mature-ovary biochemical composition did not differ between those two groups.⁴ These are measured biochemical and timing outcomes under that protocol, not a visible diagnosis that a home colony needs an outcross.

Appearance alone is not breeding value

One red-line study found females more intensely coloured at 180 days than at 90 days, while maternal colour did not predict the measured offspring colour.⁵ Background and diet studies also changed visible pigmentation without changing ancestry.⁶ A darker parent may be a strong sales specimen and still remain untested as a breeder.

A useful selection record asks two questions at once: did the candidate meet the appearance objective, and what proportion of its comparable offspring met it? Keep hatch, survival, growth and health outcomes beside the colour score so selection for appearance does not hide a deteriorating production result.

Animals outside the objective still retain their source and welfare history. Use the surplus guide before any sale, giveaway or buy-back.

Family and tank effects can be confused

A 2026 quantitative-genetics study built 75 full-sibling families nested within 37 paternal half-sibling groups and measured 1,191 offspring. Its subject was behaviour, not body colour. The authors cautioned that raising each full-sibling family in a separate tank could cause shared environment, parental effects or dominance to be mistaken for additive genetic effects.⁷ The same limitation applies to hobby colour comparisons. If family A occupies tank A and family B occupies tank B, a difference is a family-plus-tank result.

The model attributed an estimated 24.3 percent of multivariate behavioural variance to additive genetic effects, but the paper reports high uncertainty. Only about two dams per sire contributed on average, ages ranged from 41 to 215 days and all families came from one source colony. The estimate does not describe a named colour line, one breeder's personality or a selection response.

Replicate family branches across comparable systems when practical. When it is not practical, call the result a family-tank observation and reproduce it in a later generation before claiming inherited improvement.

A minimum defensible selection cycle

Our practice

1. Freeze one objective. Define the exact colour, coverage, pattern or stripe score and the age, sex, background, lighting and image method. Do not change the target after seeing the brood.
2. Preserve a foundation branch. Keep documented unmodified stock separate from the selection and outcross branches. No experimental offspring return to it.
3. Name the parents honestly. Record source, line, sex-confidence basis and family. If paternity matters, use source-isolated offspring with no unrecorded male exposure rather than assuming a purchased female's history.
4. Create a cohort at hatch. Record the hatch window and starting denominator or honest range. Keep families separate until identity no longer matters.
5. Hold the observation method constant. Grade at a defined house checkpoint and report every eligible animal, including off-pattern, translucent, missing and dead outcomes.
6. Score the whole result. Record eggs, hatch, retention, growth, health gate, phenotype distribution and number of actual contributors to the next generation.
7. Select with a written reason. Retain candidates from families that meet both the appearance objective and the operation's fitness outcomes. Do not select only the single darkest animal.
8. Repeat before renaming. One brood is a family observation. A stable line claim needs the selected result reproduced across documented generations and conditions.

The checkpoints in the cohort guide are house observation dates, not universal colour-development ages. A line-specific grading age remains provisional until its own cohorts show the phenotype can be compared consistently then.

Does the colony need new stock?

There is no validated every-year, every-five-generations or minimum-colony rule. A closed colony can contain close relatives, but colony age, animal count and a few pale offspring do not measure relatedness or prove inbreeding depression. First check known ancestry and repeated cohort outcomes.

A reversible outcross trial

1. Write the problem and the success measures before buying or moving animals.
2. Retain the foundation line unchanged as the comparison and recovery branch.
3. Quarantine the new source as its own wet contact network.
4. Assign a new experimental-cross name. Matching trade labels do not make the animals one line.
5. Compare foundation, source and crossed cohorts with the same age, grading, feeding, density, water and survival records.
6. Follow F1 and later offspring before deciding whether the cross met the stated objective. Do not sell the branch under the foundation identity while the result is unresolved.
7. Merge or rename only through a dated decision that preserves the source and cross history. A better-looking F1 does not prove lasting improvement.

New stock can also introduce organisms, treatments, identification uncertainty and a different phenotype. An outcross is therefore a genetic, biosecurity and product-identity decision at the same time, not routine maintenance.

A community record worth combining

A useful crossing record includes both parents or source populations, their confidence labels, the mating design, hatch window, every F1 and F2 denominator, common grading method and all phenotype categories. Publish the original images and the result that did not match the hoped-for chart.

This cross produced these observed categories in this cohort under this method. It does not establish universal dominance, ancestry or percentages for every seller using the same trade names.

What the evidence does not prove

• There is no validated dominance chart for the common aquarium trade lines.
• A grade or colour does not reveal genotype, parentage or genetic diversity.
• Two sellers do not automatically provide unrelated or genetically wider stock.
• A colony count does not reveal the number of effective breeding contributors.
• One full-sibling study does not establish a universal outcross schedule.
• A pale or off-pattern brood does not diagnose inbreeding.
• An attractive F1 does not prove that the result will persist in F2 or later generations.
• Candidate genes or associated variants do not identify a commercial line without validation.

Primary sources and evidence review

1. Bai et al., 2026. Decoding color variation: genome-wide characterization and expression analysis of pigment pathways in Neocaridina denticulata. BMC Genomics 27: 535.
2. Li et al., 2026. Functional Analysis of the NinaB-like Gene in Body Color Regulation of Neocaridina denticulata sinensis. BioTech 15(1): 15.
3. Tropea, Marciano and López Greco, 2022. Brothers are better than nothing: first report of incestuous mating and inbreeding depression in a freshwater decapod crustacean. Zoology 151: 125990.
4. Tropea and López Greco, 2024. To invest or not to invest, that is the question: male presence and genetic relatedness as modulators of female reproductive effort in a shrimp. Canadian Journal of Zoology 102(9): 721-734.
5. Sganga and López Greco, 2019. Assessment of potential trade-off between maternal colouration and offspring quality in the ornamental red cherry shrimp. Aquaculture Research 50(6): 1564-1573.
6. Tomas, Sganga and López Greco, 2020. Effect of background color and shelters on female pigmentation in the ornamental red cherry shrimp. Journal of the World Aquaculture Society 51(3): 775-787.
7. Wilson et al., 2026. Quantitative genetics of shy-bold behaviour and plastic response to novel predator cues in the cherry shrimp, Neocaridina davidi. Journal of Evolutionary Biology.
8. Lu et al., 2022. The occurrence process of chromatophores in three body color strains of the ornamental shrimp Neocaridina denticulata sinensis. Zoomorphology 141: 283-295.
9. Feng et al., 2025. Genome-wide identification of the crustacyanin gene family in Neocaridina denticulata sinensis based on comparative genomics and localization analysis. BMC Genomics 26: 1151.
10. Huo et al., 2025. Functional Analysis of NdBCO-like4 Gene in Pigmentation of Neocaridina denticulata sinensis. Fishes 10(3): 134.
11. Zhang et al., 2025. Role Analysis of the scarb1 Gene in the Pigmentation of Neocaridina denticulata sinensis. Animals 15(7): 901.
12. Kenny et al., 2014. Genomic Sequence and Experimental Tractability of a New Decapod Shrimp Model, Neocaridina denticulata. Marine Drugs 12(3): 1419-1437.
13. Li et al., 2022. Microinjection-based CRISPR/Cas9 mutagenesis in the decapoda crustaceans Neocaridina heteropoda and Eriocheir sinensis. Journal of Experimental Biology 225(6): jeb243702.
14. Sin et al., 2015. Identification of putative ecdysteroid and juvenile hormone pathway genes in the shrimp Neocaridina denticulata. General and Comparative Endocrinology 214: 167-176.
15. Mykles and Hui, 2015. Neocaridina denticulata: A Decapod Crustacean Model for Functional Genomics. Integrative and Comparative Biology 55(5): 891-897.
16. Huang et al., 2020. Identification of Functional SSR Markers in Freshwater Ornamental Shrimps Neocaridina denticulata Using Transcriptome Sequencing. Marine Biotechnology 22: 772-785.

Taxonomy note: molecular papers above use more than one scientific combination. We preserve each source's name and do not silently convert its result into a diagnostic claim for our maintained N. davidi lines. See the dated naming standard.