Two of the most widely used feline cell lines in biomedical research have finally given up a secret that laboratories around the world have carried, unknowingly, for decades: what the cats they came from actually looked like. The CRFK line, they report, traces back to a cat with long black fur lacking stripes and non-blue irises, while PG-4 originated from a long-furred, black-and-white bicolored cat, also without stripes and also with non-blue irises. The study, published in Companion Animal Health and Genetics, fills a gap that has persisted since the lines were first established, because no detailed record of the donors’ visible phenotypes was ever kept.
The motivation is more practical than curiosity alone. CRFK cells, derived from the kidney of a domestic cat, have served as the backbone of countless virology experiments, from feline calicivirus and feline immunodeficiency virus to feline coronavirus and endogenous retroviruses such as RD114, as well as vaccine development. PG-4 cells, established from feline fetal astrocytes, have been central to retrovirus research. Both lines have been passaged for decades in laboratories worldwide, yet the genetic background of the cats they came from, particularly traits like coat and iris color, remained undocumented. Because genes that shape coat and eye appearance are also active in many other organs, knowing which variants a cell line carries could change how results obtained from it are interpreted.
To recover the missing information, the team obtained CRFK cells (JCRB9035) and PG-4 (S+L−) cells (JCRB9125) from the JCRB Cell Bank in Osaka and subjected them to both paired-end short-read and long-read whole-genome sequencing. Short-read sequencing on an Illumina NovaSeq X Plus platform produced roughly 60 gigabases of raw data for each line, with average mapping depths of about 23.6-fold for CRFK and 23.5-fold for PG-4 against the F.catus_Fca126_mat1.0 reference genome. Long-read sequencing on Oxford Nanopore’s PromethION platform added 9.6 and 11.8 gigabases respectively, with N50 read lengths of 23.5 and 38.2 kilobases. Reads were mapped to two reference assemblies, F.catus_Fca126_mat1.0 and AnAms1.0, the latter being necessary because the former lacks a genomic region relevant to the orange coat locus.
The researchers then examined the nucleotide sequences of thirteen genes with well-established links to feline coat and iris phenotypes: TYR, TYRP1, MC1R, KIT, ARHGAP36, MLPH, ASIP, LVRN, DKK4, FGF5, KRT71, LPAR6, and PAX3. Each of these genes carries variants whose effects are known from prior genotype-phenotype work, such as TYR for colorpoint patterning, TYRP1 for brown coat, MC1R for red, KIT for white and white spotting, MLPH for dilute pigmentation, ASIP for the agouti pattern, LVRN and DKK4 for tabby variants, FGF5 for long fur, and PAX3 for blue irises. By comparing the cell line sequences against wild-type references at these variant positions, the team could read off, gene by gene, which alleles each donor cat had carried.
In CRFK cells, the analysis revealed a homozygous deletion mutation in the ASIP gene, the nonagouti allele that suppresses the striped tabby pattern; a homozygous mutation in FGF5, the long-coat allele; a homozygous blotched-tabby allele in LVRN; and a heterozygous dilute allele in MLPH. Crucially, the LVRN mutation turned out to be irrelevant here, because the ASIP mutation already eliminates striping, and the single MLPH copy has only a recessive effect. The team also checked for larger structural variants: no 95-kilobase deletion in the KIT-KDR intergenic region that produces the salmiak coat color, no 5-kilobase deletion in ARHGAP36 intron 1 that produces orange coat, no endogenous retroviral insertions in PAX3 intron 4 associated with blue irises, and no FERV1 insertion in KIT intron 1 that would produce white spotting. Long reads spanning the KIT region confirmed the absence of the insertion.
PG-4 cells told a different story. They shared the homozygous ASIP nonagouti mutation, but carried a compound heterozygous pair of FGF5 mutations, one allele with a deletion at base 474 and the other with a substitution at base 475, which together still produce long fur. They also carried heterozygous mutations in LVRN and a heterozygous Siamese-color allele in TYR, the latter recessive and therefore invisible in the donor’s appearance. The decisive finding came from long-read sequencing: six reads mapped to the KIT-FERV1 sequence, and four of them contained an FERV1 homolog insertion exceeding seven kilobases, indicating that the PG-4 genome carries a heterozygous insertion of the entire FERV1 sequence. Because the effect of this insertion on white spotting is dominant, the PG-4 donor cat would have shown white-and-black bicolored fur.
The iris color inference required additional care. Cats with white spotting can occasionally have blue or atypical eye coloration, though less commonly than cats carrying the Dominant White KIT allele. However, no reads from either cell line mapped simultaneously to PAX3 intron 4 and either the FERV1-LTR or RD114-LTR sequences, the insertions associated with blue irises in cats. On that basis, the researchers concluded that both donor cats most likely had non-blue irises. They acknowledged the inherent uncertainty in such reconstructions, presenting their results as expected or likely phenotypes rather than certainties, and illustrated the inferred appearances with example drawings of a long-haired black cat and a long-haired bicolored cat.
Beyond reconstructing the donors’ looks, the study examined where these coat and iris genes are actually active in the cat body. Mining publicly available RNA-seq data from various feline tissues in the Ensembl database, the team found that coat phenotype-related genes are expressed in many organs at levels comparable to or exceeding those in skin. TYR, TYRP1, and MLPH were highly expressed in the ear tip, retina, and optic nerve; MC1R, KIT, DKK4, and FGF5 showed high expression in the brain; ARHGAP36 was prominent in the spinal cord; ASIP and MLPH in the lungs; and KIT, DKK4, and LPAR6 were expressed throughout the body, with notable activity in reproductive organs. PAX3, the iris color gene, was highly expressed in the cerebellum as well as visual organs, and, along with DKK4, in kidney and embryonal tissue.
This tissue-wide expression raises a question the authors consider important: variants of genes chosen for their visible effects may also influence physiological functions far from the skin and eyes. The researchers note that such genetic effects are likely weaker than those of the rearing environment and may be insignificant in young, healthy animals, but as domestic cats live longer, individual genotypes could exert increasingly apparent effects on disease susceptibility in specific organs. They also caution that results from cultured cells must be interpreted with both the experimental findings and the genotype and expression profiles of the lines in mind, information that simply has not existed until now even for a line as heavily used as CRFK. Detailed genomic data would also allow genome editing to create derivative lines differing only in specific traits, enabling direct tests of, for example, the role of a novel KIT isoform produced by FERV1-LTR insertion in white coat formation, and potentially shedding light on deafness in white-furred, blue-eyed cats and on how melanocyte destruction from KIT mutations affects the nervous system.
The authors are explicit about the limitations. Their findings apply to the specific cell stocks they sequenced; chromosome numbers in CRFK have been reported to fluctuate between 36 and 38, and the unreported karyotype of PG-4 likely differs from the normal cat genome, so other laboratories’ stocks may behave slightly differently. Detecting large structural variations was difficult with roughly 60 gigabases of paired-end data and only about 10 gigabases of long-read data, and repeat sequences and endogenous retroviral homologs remain incompletely resolved. Preliminary analyses of MHC regions and mitochondrial DNA variants were deposited but could not identify the donor breeds, a goal that may require deeper long-read sequencing. The raw sequence data have been deposited in the DDBJ Sequence Read Archive under BioProject PRJDB37530, making this genetic portrait of two anonymous donor cats available to anyone whose experiments depend on the cells they left behind.
Subject of Research: Whole-genome sequencing of feline cell lines to infer the coat and iris phenotypes of their donor cats
Article Title: Whole-genome sequencing of CRFK and PG-4 cells to infer the phenotype of the original donor cats
Article References: Tanaka, G., Goto, R., Komoto, T., Kubota, A., Hayashi, R., Igawa, T., Sakamoto, N., & Awazu, A. (2026). Whole-genome sequencing of CRFK and PG-4 cells to infer the phenotype of the original donor cats. Companion Animal Health and Genetics, 13(1), Article 1. https://doi.org/10.1186/s40575-026-00150-9
Image Credits: AI Generated
DOI: 10.1186/s40575-026-00150-9
Keywords: CRFK cells, PG-4 cells, whole-genome sequencing, domestic cat, coat color genetics, iris color, FGF5, KIT, PAX3, cell lines, RNA-seq, Felis catus
News Source: Juliet Wilcox. (October 29, 2026). Research report: Whole-genome sequencing of CRFK and PG-4 cells to infer the phenotype of the original donor cats. Scienmag.
