Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 May 1.
Published in final edited form as: Dev Dyn. 2016 Mar 6;245(5):605–613. doi: 10.1002/dvdy.24398

Automated, high-throughput, in vivo analysis of visual function using the zebrafish

C Anthony Scott 1, Autumn N Marsden 2, Diane C Slusarski 1
PMCID: PMC4844763  NIHMSID: NIHMS761762  PMID: 26890697

Abstract

Background

Modern genomics has enabled the identification of an unprecedented number of genetic variants, which in many cases are extremely rare, associated with blinding disorders. A significant challenge will be determining the pathophysiology of each new variant. The zebrafish is an excellent model for the study of inherited diseases of the eye. By 5 days-post-fertilization (dpf) they have quantifiable behavioral responses to visual stimuli. However, visual behavior assays can take several hours to perform or can only be assessed one fish at a time.

Results

To increase the throughput for vision assays, we used the Viewpoint Zebrabox to automate the visual startle response and created software, Visual Interrogation of Zebrafish Manipulations (VIZN), to automate data analysis. This process allows 96 zebrafish larvae to be tested and resultant data to be analyzed in under 35 minutes. We validated this system by disrupting function of a gene necessary for photoreceptor differentiation and observing decreased response to visual stimuli.

Conclusions

This automated method along with VIZN allows rapid, high-throughput, in vivo testing of zebrafish’s ability to respond to light/dark stimuli. This allows the rapid analysis of novel genes involved in visual function by morpholino, CRISPRS, or small molecule drug screens.

Keywords: vision, behavior tracking, visual assay, blinding disorders

Graphical abstract

We automated a visual behavioral assay and created software to automate the data analysis.

graphic file with name nihms761762u1.jpg

Introduction

Eye disease and visual impairment coupled with age-related vision loss poses a significant public health concern. Advances in personal genomic and bioinformatic approaches have accelerated the identification of potential disease-associated genes, many that would have otherwise never been implicated in visual function. However, such identified alleles are typically extremely rare in the population and standard genetic methods cannot be applied to verify if the candidate allele is disease-associated. While the mouse model has contributed much in the understanding of disease, it is not well-suited for rapid evaluation of disease-candidate genes (Young, 1985; Bassett and Wallace, 2012). The zebrafish, by virtue of large clutches of offspring and speed of development, facilitate high-throughput in vivo screens to directly study human diseases (Dooley and Zon, 2000; Kari et al., 2007; Lieschke and Currie, 2007).

The zebrafish visual system is more similar to human with respect to rod/cone utilization, a result of also being diurnal. This is in contrast with rod-dominant murine models, which are nocturnal and have more rods to facilitate higher acuity night-vision (Wikler and Rakic, 1990; Menger et al., 2005; Fleisch and Neuhauss, 2006). By 5 days-post-fertilization (dpf) the zebrafish eye is arranged, or laminated, in the same five distinct layers consisting of all the same cell types present in human eyes (Chhetri et al., 2014). Mouse models show the same lamination by 20 days of gestation, but functional vision will not be completely developed until more than one month after birth, compared to the fully functioning zebrafish eye at 5dpf (Sorsby et al., 1954; Easter Jr and Nicola, 1996; Kitambi et al., 2011).

Zebrafish show characteristic behavioral responses to visual stimuli. While several assays have been developed to test vision in zebrafish, there are two that use stereotypic swimming behavior: the visual-startle-response (VSR) and visual-motor-response (VMR), both of which are amenable for high throughput analysis. In VSR, 5dpf larva respond to a sudden change of light by rapidly changing direction and swimming away from the perceived threat (Easter Jr and Nicola, 1996). VMR tracks the spike of locomotion (startle response) at light transitions and also evaluates the time to return to baseline activity (Emran et al., 2008). VMR assays typically acclimate the 5dpf larvae for 1–3 hours and involve multiple rounds of lights on and off conditions lasting 30 minutes each (Emran et al., 2008; Maurer et al., 2010; Maurer et al., 2011). In VSR, a visual startle stimulus is induced by performing rapid changes in light intensity using a shutter located between the light source and the animal on a microscope stage (Fig. 1A). An abrupt movement within one second of visual stimuli application is scored as a positive response (Fig 1B), and each larva is subjected to five trials spaced approximately 30 seconds apart. The VSR assay is highly flexible in that different number of startle stimuli have been successfully employed, typically in the range of five to ten (Easter Jr and Nicola, 1996; Burgess and Granato, 2007). The VSR has proven to be a powerful assay to implicate gene function in vision and for testing disease associated alleles; however, the VSR assay is typically performed one larvae at a time, making it time consuming (Baye et al., 2011; Pretorius et al., 2011; DeLuca et al., 2015).

Fig.1.

Fig.1

Open in a new tab

Performing the VSR. Photos showing how the manual VSR (A–A′) and the automated VSR are setup (C). The characteristic C-bend that occurs during the VSR can be detected when viewing a recording of the experiment frame-by-frame (B).

Some visual assays, such as the VMR, have been successfully automated (Zhang et al., 2012; Gao et al., 2014; Spulber et al., 2014). However, a VMR assay can take several hours to perform, limiting the amount of fish that can be tested in one day. The VSR assay, by contrast, takes only a few minutes per fish to perform, but can take multiple hours to achieve high numbers. Our lab has previously adapted Easter and Nicola (1996) VSR method to analyze visual function (Pretorius et al., 2010; Baye et al., 2011; Pretorius et al., 2011)(experimental procedures). To this end, we adapted the Zebrabox (Viewpoint Live Sciences, Fig. 1C) motion tracking system to perform the VSR assay in a 96-well plate in less than 33 minutes (Fig. 2). This automation drastically reduces the time is takes to analyze large numbers of fish, but leads to the output of very large datasets. To manage the unwieldy spreadsheets, we developed software, Visual Interrogation of Zebrafish Manipulations (VIZN) that automatically processes the data in a matter of seconds. It implements a graphical user interface (GUI) that allows fast and easy visualization of the data, including graphs, individual response values, and statistics. We validated the system by testing visually compromised zebrafish and comparing manual scoring to automated analysis. The Zebrabox combined with VIZN provides a new automated system that allows for high throughput visual analysis of genetic manipulations (using gene knockdown/knockout techniques) and could be used for small molecule drug assays to test possible therapeutic effects in visually-defective strains (Zon and Peterson, 2005; Kari et al., 2007).

Fig. 2.

Fig. 2

Open in a new tab

Zebrabox experimental design. Diagram of the Zebrabox protocol showing the 30 minute adaptation phase and 2.5 min testing phase (A). Plot of the average activity levels from one row of wildtype fish (B). Zoomed in region of the dotted box showing the clear startle response peaks (C).

Results and Discussion

The Zebrabox automates the VSR and VIZN crunches the data

To automate the VSR, we first created a behavior tracking protocol that could monitor behavior in multiwall plates in the Zebrabox. We used the Viewpoint Zebralab quantization software to record the activity of the individual larva in a 96-well plate in time bins of 1 second. The first 30 minutes are constant light to capture the background/baseline locomotor activity of the larvae. We have found from previous tracking analyses that 30 minutes baseline activity in the Zebrabox produces reliable behavior from experiment to experiment (Mei et al., 2013). The testing phase mimics the manual protocol of 1 second of darkness followed by 29 seconds of light, and this is repeated 5 times (Fig. 2A). When we run a 96-well plate in the Zebrabox with our described automated VSR protocol, we are returned a 31 megabyte spreadsheet with over 187,000 rows of data in 21 columns that takes significant time to manually process and introduces the potential for human error. To accelerate this analysis, we developed VIZN to transform the raw spreadsheet into useable graphs and tables.

VIZN software is hosted on a public server, allowing anyone to upload and process raw data. To process the data, VIZN averages the “actinteg” (activity integration) value, which is the overall quantification of pixel movement captures by the camera for each larva during each 1 second time window. The 30 minute adaptation period is used to calculate the baseline movement for each row of larvae. All activity levels in the experiment are divided by this baseline to normalize the actinteg values to approximately 1 (Fig. 2B). Next, VIZN looks at the 1 second periods that occur during the 5 flashes of darkness and decides if the activity at these points are higher than baseline by a user input threshold (Fig. 2C). The program is designed to allow flexibility in regards to the activity by allowing the user to input their own threshold. This threshold defaults to 2.5, meaning the activity level at this time point needs to be 2.5X higher than the baseline to be considered a positive response. If it is higher, VIZN counts this as a response. VIZN performs this analysis for all 5 time points for every well of the plate. In our analyses of 5dpf larva at 28°C, we leave the threshold at default. However, including the option to adjust thresholds allow users flexibility to compensate for differing ages, temperatures, or genetic backgrounds which may have different activity profiles.

VIZN was programmed using the R programming language. We developed it to be extremely user friendly using the Shiny package to develop the GUI. This GUI allows the user to group rows together in any order or combination and perform basic statistics (Fig. 3A). In this manner, multiple rows of a similar manipulation can be grouped together, but they do not need to be loaded next to each other on the physical plate. The GUI also has a feature to examine individual responses (Fig. 3B). The software automatically excludes empty wells or wells containing larva that did not move during the entire experiment (noted as NA) to differentiate them from wells in which the larva was motile but the number of VSRs was zero. In any given plate from wildtype clutches, 3% on average are excluded (with generally 0–1 non-motile fish in any given row). In our manual assay, we test for touch-response to select embryos which have the ability to swim. Although to allow for high through put analysis, we eliminated that step. Another approach to reduce the number excluded would be to immediately re-run the plate in the zebrabox. In one example, total non-motile dropped from 5/96 to 1/96 in the second run (data not shown). VIZN was developed to assist investigators in handling spreadsheets without having to learn the underlying programming language. VIZN is presently hosted publicly online at the following URL where it continues to be updated: https://anthony-scott.shinyapps.io/vision_web_app. The source code is available upon request.

Fig. 3.

Fig. 3

Open in a new tab

Graphical User Interface (GUI). We developed the GUI to be user friendly. It can help group rows and analyze the results (A) and instantly generate a table of all responses by well (B). This data can then be exported and analyzed in any other graphing or statistics program desired.

VIZN is consistent with human judgement of the VSR

We next set out to validate that VIZN can mimic the human decisions made when the VSR is performed manually. To establish a baseline we performed the VSR manually, one embryo at a time, on three different zebrafish clutches on three different days. The average number of responses of all manually tested larvae was 3.3/5 (Fig. 4A). When we performed the VSR automated assay on a similar number of fish and analyzed these results using VIZN, we obtained an average of 3.5/5 responses. This result is not significantly different from the manually performed VSR (Fig. 4A). This demonstrates that the VSR is a consistent, reproducible assay that can be successfully automated using the Zebrabox and analyzed using VIZN. When the video recording of the experiment is analyzed frame-by-frame, the larval zebrafish can be observed performing the characteristic C-bend that occurs with the manual VSR during the dark flash (Fig. 1B, Supplemental Movie 1). This further supports accuracy of our automated method.

Fig. 4.

Fig. 4

Open in a new tab

Manual and automated VSR. Three VSR analyses were performed manually and one with equivalent numbers run in the Zebrabox and analyzed with VIZN (A). No significant difference was observed between groups, box-and-whisker plot representing min, max, and quartiles.

The VIZN assay is accurate and reproducible between different clutches of embryos

To further verify the consistency and reproducibility and the VIZN assay we performed the VSR on groups of wildtype larvae. First, we wanted to verify camera in the Zebrabox has uniform detection of the zebrafish across all 96 wells. We preformed the VIZN assay on a plate of 96 wildtype 5dpf zebrafish and averaged the responses by row. When comparing the groups by one-way ANOVA, no significant difference is observed (Fig. 5A). Next, we compared the response between three different clutches of embryos born and tested on different days. When comparing all three groups by one-way ANOVA we see no significant difference between the groups (Fig. 5B). Together these results indicate that the Zebrabox shows no bias between rows and the number of responses are consistent between days.

Fig. 5.

Fig. 5

Open in a new tab

The VIZN analysis of Zerabox VSR data is consistent. A full 96 well plate of wildtype 5dpf zebrafish was analyzed. One-way ANOVA shows no significant difference between any row on the plate (A). One-way ANOVA comparing different clutches run on different days shows no significant difference (B), box-and-whisker plot representing min, max, and quartiles.

VIZN shows a decreased VSR in cone-rod-homeobox (CRX) knockdown zebrafish

To evaluate the ability of VIZN to detect visually defective larva, we performed gene knockdown of CRX, a known transcription factor required for photoreceptor differentiation using a validated Morpholino (MO) (Liu et al., 2001; Shen and Raymond, 2004). We have previously demonstrated reduced VSR in CRX MO-injected (morphant) larvae (Nishimura et al., 2010; Pretorius et al., 2010; Baye et al., 2011; Pretorius et al., 2011). At 5dpf CRX morphants show normal morphology (Fig. 6A–B). Wildtype larvae and CRX morphant larvae were loaded into a 96 well plate and tested for VSR using the Zebrabox and analyzed with VIZN. Results show that CRX morphants respond significantly less than wildtype (p < 0.05, Fig. 6C), indicating that CRX morphants have a visual defect, successfully detected by VIZN.

Fig. 6.

Fig. 6

Open in a new tab

Knockdown of cone-rod homeobox shows VSR defects. Wildtype (A) and CRX morphant (B) larvae were subjected to the automated VSR and analyzed with VIZN. CRX morphant larvae responded less than WT to the visual stimuli (C), box-and-whisker plot representing min, max, and quartiles, t-test p-value < 0.05.

Traditional and modern zebrafish techniques combined with VIZN creates a powerful, high-throughput method to test novel vision-associated genes. Of note, the robustness of the defect in visual response may vary with the targeted gene of interest, and this is an important consideration to keep in mind during VIZN experiment design. If a strong visual defect is suspected an experiment may be able to be performed with two rows (n=24) of fish. However if a more subtle defect is suspected, it will be necessary to only perform control and treatment by loading four rows (n=48) of each. For example, at the dose of CRX MO we are using we observe a subtle visual defect. A power analysis performed with our expected results recommended a sample size of 33 to achieve appropriate power. Moreover, multiple plates can be run and datasets can be combined by exporting data from VIZN using the results tab (Fig. 3B). Because both automated and manual analyses make discrete calls in whole numbers, not in smaller increments, data sets appear to have large standard deviations, and should be taken into account in experimental design and statistical analysis.

VIZN differentiates responses from background movements

Some vision mutants or morphants can have coloboma or other gross morphological defects. To test these extreme conditions we generated “eyeless larvae” that still have the ability to swim, and still maintain a touch response. We used this to test VIZN’s ability to accurately differentiate the VSR against spontaneous background movement from embryos that cannot respond to visual stimuli. To generate larvae without eyes, we manipulated Wnt signaling by overexpressing wnt8, which leads to a loss of anterior neural structures including the forebrain and eyes (Kelly et al., 1995; Fredieu et al., 1997). Although larvae overexpressing wnt8 lack eyes and a forebrain, they still swim spontaneously and respond to touch, thus any VSR counts detected would be erroneous calls by the software (Fig. 7A–B). VIZN analysis of the data of wildtype larva shows an overall average number of response of 3.2/5 (Fig. 7C). The larvae lacking eyes show an average response number of 0.4/5, demonstrating VIZNs ability to correctly differentiate between VSR movement and spontaneous background movement. Due to the optics of the sides of the wells, it is unlikely that individual larvae can visualize neighbors. In addition, 5dpf zebrafish larvae do not yet exhibit schooling behavior and therefore any individual swimming behavior would not influence neighboring larvae (Neuhauss et al., 1999).

Fig. 7.

Fig. 7

Open in a new tab

Eyeless embryos do not respond to VSR. By 2dpf uninjected embryos have begun to develop eyes (A). Wnt8 injection prevents proper development of these structures (B). These larvae were raised to 5dpf and analyzed with VIZN along with wildtype siblings. These “eyeless” larva have an average VSR of 0.4 indicating VIZN does not detect spontaneous movements as responses (C), box-and-whisker plot showing min, max, and quartiles, t-test p-value <0.0001.

Development in near-constant darkness does not affect visual function

Now that we have a high-throughput method of testing the VSR in fish, we wanted to test the effect of rearing larvae in standard lab environments. It has been reported that rearing zebrafish in constant light or constant darkness can impair their visual function (Bilotta, 2000; Fleisch and Neuhauss, 2006). Because most standard laboratory incubators do not have light cycling, we sought to determine if rearing embryos in a standard incubator was detrimental to visual function. We separated siblings at 2dpf with half placed in a light-controlled incubator simulating normal day/night cycling conditions and the other half in a standard incubator without any light. As described in the experimental methods both groups were light adapted for 2 hours prior to transferring to a 96 well plate and VIZN analysis. We observed no significant difference in the visual startle responses between the two groups (Fig. 8). Therefore, any gross visual deficiency that may occur by rearing in darkness is overcome by light adaptation prior to visual testing.

Fig. 8.

Fig. 8

Open in a new tab

Near-constant darkness does not affect the VSR. Siblings were reared in a day/night cycling incubator or a standard incubator with no lights. No significant difference was measured between groups (A), box-and-whisker plot representing min, max, and quartiles.

VIZN can be a useful pre-screen of a larval clutch

This automated VSR assay allows a much greater number of fish to be analyzed in one day than the manual method. Assuming 96 larvae are analyzed at a time for 33 minutes each, it is conceivable to test over 1000 larvae in one day, easily allowing multiple biological replicates to be performed. Because of the short time it takes to test one plate of fish, you can also re-run the sample plate of fish multiple times in the Zebrabox producing multiple technical replicates. By performing multiple technical replicates on the same plate of fish, VIZN can be used to identify fish that maintain zero responses to the visual stimuli after multiple runs in the Zebrabox. This could serve as a helpful pre-screen of fish that could then be isolated for further experiments.

Experimental Procedures

Animal Care

Zebrafish are maintained in standard conditions under the approval of the University of Iowa IACUC. Embryos are collected from natural spawning and raised between 28.5°C and 30°C with no more than 50 embryos per 100mm plate. Embryo plates are cleared of dead every day with water changes occurring as needed.

Manual Vision Startle Response

VSR has been adapted from (Easter Jr and Nicola, 1996). 5dpf larvae are light-adapted in a glass-door incubator maintained at 28°C for at least 1 hour. Individual embryos are placed into a small plate and placed under the microscope. Larvae are subjected to 5 flashes of darkness, using the microscopes shutter switch, each separated by 30 seconds of light. Larvae are monitored during each flash for a VSR.

Automated Startle Response

Prior to testing, 5dpf larvae are cultured in a light environment for 2 hours, then transferred to 96-well plate and placed into the Zebrabox. The fish were subjected to the previously described protocol during motion tracking. The software records overall activity of the zebrafish in 1-second time bins for the 32.5 minute duration of the experiment. The Zebrabox lighting unit is capable of emitting 8000 Lux of light at 550nm when set at maximum. For our experiments we ran the lighting unit at 12% power. The inside of the Zebrabox is constantly illuminated by an infra-red light allowing the camera to record and detect movements in darkness. We set the detection sensitivity level in the Zebrabox capture software to 20. This focuses the movement detection on the head of each larvae. The equipped camera is capable of recording at 60 frames per second.

Microinjection

At the 1–2 cell stage of development embryos were injected with 8 nanograms of a translation blocking CRX morpholino, or 4 picograms of Wnt8 mRNA. Mopholinos were ordered from gene-tools.

CRX MO Sequence: 5′-GGCTGCTTTATGTAGGACATCATTC-3′

Statistical Analysis

Data was exported from VIZN using the Download Data button from the Results Table tab. The response of an individual larvae is represented as a whole number between 0–5. Larvae responses from similar treatments were grouped together and imported into GraphPad Prism 6 for statistical analyses and generation of the box-and-whisker plots. Average number of responses, standard deviation, and standard error were calculated for each group. When comparing between two groups of larvae (Figure 6, 7, 8) an unpaired t-test was performed. When comparing between three or more groups (Figure 4, 5) an ordinary one-way ANOVA was used. P-values of less than < 0.05 were considered significant for both tests.

Supplementary Material

Supp VideoS1
Download video file (1.6MB, m2v)

Bullet Points.

Automated the visual startle response in zebrafish, Created free software to instantaneously analyze the large spreadsheet of generated data, Creates opportunity for rapid, high-throughput testing of novel genes which may be implicated in blinding disorders

Acknowledgments

Thank you to all the Slusarski lab members for comments on the manuscript. C.A.S and A.N.M. are supported by American Heart Association pre-doctoral fellowships.

Grant Sponsor: NIH – R01 EY011298, R01 EY07168

References

  1. Bassett EA, Wallace VA. Cell fate determination in the vertebrate retina. Trends in neurosciences. 2012;35:565–573. doi: 10.1016/j.tins.2012.05.004. [DOI] [PubMed] [Google Scholar]
  2. Baye LM, Patrinostro X, Swaminathan S, Beck JS, Zhang Y, Stone EM, Sheffield VC, Slusarski DC. The N-terminal region of centrosomal protein 290 (CEP290) restores vision in a zebrafish model of human blindness. Human molecular genetics. 2011;20:1467–1477. doi: 10.1093/hmg/ddr025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bilotta J. Effects of abnormal lighting on the development of zebrafish visual behavior. Behavioural brain research. 2000;116:81–87. doi: 10.1016/s0166-4328(00)00264-3. [DOI] [PubMed] [Google Scholar]
  4. Burgess HA, Granato M. Modulation of locomotor activity in larval zebrafish during light adaptation. Journal of Experimental Biology. 2007;210:2526–2539. doi: 10.1242/jeb.003939. [DOI] [PubMed] [Google Scholar]
  5. Chhetri J, Jacobson G, Gueven N. Zebrafish—on the move towards ophthalmological research. Eye. 2014;28:367–380. doi: 10.1038/eye.2014.19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. DeLuca AP, Whitmore SS, Barnes J, Sharma TP, Westfall TA, Scott CA, Weed MC, Wiley JS, Wiley LA, Johnston RM. Hypomorphic mutations in TRNT1 cause retinitis pigmentosa with erythrocytic microcytosis. Human molecular genetics. 2015:ddv446. doi: 10.1093/hmg/ddv446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dooley K, Zon LI. Zebrafish: a model system for the study of human disease. Current opinion in genetics & development. 2000;10:252–256. doi: 10.1016/s0959-437x(00)00074-5. [DOI] [PubMed] [Google Scholar]
  8. Easter SS, Jr, Nicola GN. The development of vision in the zebrafish (Danio rerio) Developmental biology. 1996;180:646–663. doi: 10.1006/dbio.1996.0335. [DOI] [PubMed] [Google Scholar]
  9. Emran F, Rihel J, Dowling JE. A behavioral assay to measure responsiveness of zebrafish to changes in light intensities. Journal of visualized experiments: JoVE. 2008 doi: 10.3791/923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fleisch VC, Neuhauss SC. Visual behavior in zebrafish. Zebrafish. 2006;3:191–201. doi: 10.1089/zeb.2006.3.191. [DOI] [PubMed] [Google Scholar]
  11. Fredieu JR, Cui Y, Maier D, Danilchik MV, Christian JL. Xwnt-8 and lithium can act upon either dorsal mesodermal or neurectodermal cells to cause a loss of forebrain in Xenopus embryos. Developmental biology. 1997;186:100–114. doi: 10.1006/dbio.1997.8566. [DOI] [PubMed] [Google Scholar]
  12. Gao Y, Chan RH, Chow TW, Zhang L, Bonilla S, Pang C-P, Zhang M, Leung YF. A high-throughput zebrafish screening method for visual mutants by light-induced locomotor response. Computational Biology and Bioinformatics, IEEE/ACM Transactions on. 2014;11:693–701. doi: 10.1109/TCBB.2014.2306829. [DOI] [PubMed] [Google Scholar]
  13. Kari G, Rodeck U, Dicker AP. Zebrafish: an emerging model system for human disease and drug discovery. Clinical Pharmacology & Therapeutics. 2007;82:70–80. doi: 10.1038/sj.clpt.6100223. [DOI] [PubMed] [Google Scholar]
  14. Kelly GM, Greenstein P, Erezyilmaz DF, Moon RT. Zebrafish wnt8 and wnt8b share a common activity but are involved in distinct developmental pathways. Development. 1995;121:1787–1799. doi: 10.1242/dev.121.6.1787. [DOI] [PubMed] [Google Scholar]
  15. Kitambi SS, Chandrasekar G, Addanki VK. Teleost fish—a powerful models for studying development, function and diseases of the human eye. Curr Sci. 2011;100:1815. [Google Scholar]
  16. Lieschke GJ, Currie PD. Animal models of human disease: zebrafish swim into view. Nat Rev Genet. 2007;8:353–367. doi: 10.1038/nrg2091. [DOI] [PubMed] [Google Scholar]
  17. Liu Y, Shen Y-C, Rest JS, Raymond PA, Zack DJ. Isolation and characterization of a zebrafish homologue of the cone rod homeobox gene. Investigative Ophthalmology and Visual Science. 2001;42:481–487. [PubMed] [Google Scholar]
  18. Maurer CM, Huang Y-Y, Neuhauss SC. Application of zebrafish oculomotor behavior to model human disorders. Reviews in the Neurosciences. 2011;22:5–16. doi: 10.1515/RNS.2011.003. [DOI] [PubMed] [Google Scholar]
  19. Maurer CM, Schönthaler HB, Mueller KP, Neuhauss SC. Distinct retinal deficits in a zebrafish pyruvate dehydrogenase-deficient mutant. The Journal of Neuroscience. 2010;30:11962–11972. doi: 10.1523/JNEUROSCI.2848-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mei X, Wu S, Bassuk AG, Slusarski DC. Mechanisms of prickle1a function in zebrafish epilepsy and retinal neurogenesis. Disease models & mechanisms. 2013;6:679–688. doi: 10.1242/dmm.010793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Menger GJ, Koke JR, Cahill GM. Diurnal and circadian retinomotor movements in zebrafish. Visual neuroscience. 2005;22:203–209. doi: 10.1017/S0952523805222083. [DOI] [PubMed] [Google Scholar]
  22. Neuhauss SC, Biehlmaier O, Seeliger MW, Das T, Kohler K, Harris WA, Baier H. Genetic disorders of vision revealed by a behavioral screen of 400 essential loci in zebrafish. The Journal of neuroscience. 1999;19:8603–8615. doi: 10.1523/JNEUROSCI.19-19-08603.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Nishimura DY, Baye LM, Perveen R, Searby CC, Avila-Fernandez A, Pereiro I, Ayuso C, Valverde D, Bishop PN, Manson FD. Discovery and functional analysis of a retinitis pigmentosa gene, C2ORF71. The American Journal of Human Genetics. 2010;86:686–695. doi: 10.1016/j.ajhg.2010.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Pretorius PR, Aldahmesh MA, Alkuraya FS, Sheffield VC, Slusarski DC. Functional analysis of BBS3 A89V that results in non-syndromic retinal degeneration. Human molecular genetics. 2011;20:1625–1632. doi: 10.1093/hmg/ddr039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pretorius PR, Baye LM, Nishimura DY, Searby CC, Bugge K, Yang B, Mullins RF, Stone EM, Sheffield VC, Slusarski DC. Identification and functional analysis of the vision-specific BBS3 (ARL6) long isoform. PLoS genetics. 2010;6:e1000884. doi: 10.1371/journal.pgen.1000884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Shen Y-c, Raymond PA. Zebrafish cone-rod (crx) homeobox gene promotes retinogenesis. Developmental biology. 2004;269:237–251. doi: 10.1016/j.ydbio.2004.01.037. [DOI] [PubMed] [Google Scholar]
  27. Sorsby A, Koller P, Attfield M, Davey J, Lucas D. Retinal dystrophy in the mouse: histological and genetic aspects. Journal of Experimental Zoology. 1954;125:171–197. [Google Scholar]
  28. Spulber S, Kilian P, Wan Ibrahim W, Onishchenko N, Ulhaq M, Norrgren L, Negri S, Di Tuccio M, Ceccatelli S. PFOS induces behavioral alterations, including spontaneous hyperactivity that is corrected by dexamfetamine in zebrafish larvae. PloS one. 2014;9:e94227. doi: 10.1371/journal.pone.0094227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Wikler KC, Rakic P. Distribution of photoreceptor subtypes in the retina of diurnal and nocturnal primates. The Journal of Neuroscience. 1990;10:3390–3401. doi: 10.1523/JNEUROSCI.10-10-03390.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Young RW. Cell differentiation in the retina of the mouse. The Anatomical Record. 1985;212:199–205. doi: 10.1002/ar.1092120215. [DOI] [PubMed] [Google Scholar]
  31. Zhang L, Chong L, Cho J, Liao P-C, Shen F, Leung YF. Drug Screening to Treat Early-Onset Eye Diseases: Can Zebrafish Expedite the Discovery? The Asia-Pacific Journal of Ophthalmology. 2012;1:374–383. doi: 10.1097/APO.0b013e31827a9969. [DOI] [PubMed] [Google Scholar]
  32. Zon LI, Peterson RT. In vivo drug discovery in the zebrafish. Nature reviews Drug discovery. 2005;4:35–44. doi: 10.1038/nrd1606. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp VideoS1
Download video file (1.6MB, m2v)

RESOURCES