Aging is a shared risk factor for many eye diseases, including age-related macular degeneration, glaucoma, cataract and diabetic retinopathy. Traditional model organisms such as mice and zebrafish live too long, which makes age-related changes hard to study. The African killifish (Nothobranchius furzeri) is the shortest-lived vertebrate that breeds in the lab (4-6 months), and it shows many hallmarks of aging, including DNA damage, cellular senescence and declining visual acuity. Yet existing genomic resources for the killifish are limited: the old Nfu_20140520 reference is incomplete, and the newer NfurGRZ-RIMD1, though more complete, comes with a poorly annotated transcriptome (nearly half of its genes are labeled 'LOC###'). This hampers gene expression analysis and integration with epigenetic data. The present work fills that gap by building an improved retinal transcriptome from long RNA reads.
Retinas from young (8-week-old) and old (22-week-old) killifish were used to extract total RNA. Full-length cDNA libraries were prepared with the PacBio Kinnex protocol and then sequenced on the PacBio Revio platform. The resulting long reads were processed with the Iso-Seq pipeline: demultiplexing, identification of full-length non-chimeric transcripts, clustering and isoform collapsing. The transcripts were then classified and filtered with SQANTI3, using 37 published bulk retinal RNA-seq datasets to validate splice sites. Open reading frames were predicted with TransDecoder, and functional annotation was done with EggNOG-mapper using orthology to teleost fish. To test the usefulness of the new transcriptome, published bulk RNA-seq data (14 tissues) and single-cell RNA-seq data (retina at 6, 12 and 18 weeks) were reanalyzed, comparing mapping against three references: Nfu_20140520, NfurGRZ-RIMD1 and the new PacBio transcriptome.
The new transcriptome contains 21,014 genes and 165,015 transcripts, of which more than 53% are novel isoforms. The share of genes with biologically meaningful names rose from 52.9% for NfurGRZ-RIMD1 to 71.9%, while the share of 'LOC###' dropped from 47.1% to less than 1%. On remapping of bulk RNA-seq, the share of mapped reads increased for most tissues, especially neural ones (retina, brain, spinal cord), but decreased for liver, fat and intestine. In single-cell RNA-seq, the share of mapped reads rose by 6-8% compared with NfurGRZ-RIMD1 and by 10-12% compared with Nfu_20140520, and the number of detected cells increased by 24-26%. Differential expression analysis between 21- and 7-week-old retinas identified 1,683 genes, of which 80.2% have human ortholog names. Among them, the previously validated genes optn, nfkb2 and pcdh15 were confirmed. Trajectory analysis of expression revealed clusters linked to aging: decreased expression of mitochondrial and translational genes, increased expression of inflammation, angiogenesis and extracellular matrix remodeling genes. The new transcriptome also allowed correct annotation of genes that had been missing before, including crx, nr2e3, rho and others. An age-related loss of expression was found for one of the two tandemly duplicated green opsin genes rh2 (PB.6621), confirmed by in situ hybridization chain reaction.
The work delivers an improved killifish retinal transcriptome that increases the accuracy of gene expression analysis during aging. Compared with previous resources, the new transcriptome contains more full-length isoforms and functional annotations, which enables a more complete pathway analysis and biological interpretation. Unlike a recent study of the telencephalon, this work uses the new chromosomal assembly NfurGRZ-RIMD1 and includes a direct comparison of three references on the same data, demonstrating the advantages of the new resource. This makes the killifish a more powerful model for studying age-related retinal diseases.
The next task will be to build a consensus transcriptome useful for all killifish tissues, since the current resource is optimized for the retina and nervous system. Integration with epigenetic data is also needed to study the mechanisms of aging. Better annotation of the genome and transcriptome will open the door to studies at the level of isoforms and splicing regulation during aging.
The new transcriptome will improve research on retinal aging and neurodegeneration in the killifish model. It may also be useful for comparative fish genomics and for studying the evolution of visual pigments.
The next step will be to extend the transcriptome to other tissues and create a single reference for the whole organism. There are also plans to use the new resource for a detailed analysis of age-related changes in splicing and isoforms in the retina.
The work solves the problem of incomplete genome annotation in a new model system, which had limited the interpretation of transcriptomic data. This is directly tied to unresolved questions about the molecular mechanisms of age-related retinal diseases such as age-related macular degeneration, where changes in gene expression play a key role.
The African killifish lives only 4-6 months, yet its retina ages much like a human one: DNA damage accumulates and visual acuity declines. It also carries two nearly identical green opsin genes, and with age one of them goes silent.
In Alan Dean Foster's novel 'The Man Who Used the Universe,' aliens live very short lives but have complex biology. The killifish, as a model of accelerated aging, resembles such fictional species, letting us study processes that stretch over decades in humans.
Not published on arXiv. Here are its page, the PDF and the formalization: you can check it yourself, not only read our account.