Poster #75: Shamdeed Kabir

Extending the chicken breast muscle transcriptome through high-confidence novel isoform discovery using long-read sequencing

Shamdeed Kabir, BA (PhD student in Bioinformatics Data Science, University of Delaware); Alexandra Cooper, BS (PhD student in Animal and Food Sciences, University of Delaware); Hailey Hoffman, BS (PhD student in Animal and Food Sciences, University of Delaware); Behnam Abasht, PhD (Professor of Animal Genetics, Department of Animal and Food Sciences, University of Delaware)

Long-read sequencing has recently enabled direct identification of novel transcript isoforms beyond reference annotations in chicken, as demonstrated by a FAANG consortium study of 68 samples across 19 tissues reporting 46,175 novel transcripts using a total of 23.8 million nanopore reads. Subsequently, the Gallus Enriched Gene Annotation (GEGA) database was created that integrates NCBI, Ensembl, FAANG, and NONCODE resources. In this study, we have performed PCR-cDNA nanopore sequencing of 191 Pectoralis (p.) major muscle samples collected to investigate the transcriptomic basis of Wooden Breast (WB) myopathy and its histopathological sub-phenotypes. For accurate characterization and quantification of the isoform-resolved transcriptome, a well-annotated genome is crucial. As a discovery pilot, we pooled a total of 12 samples with 4 samples (2 males; 2 females) across 3 hatch groups having 6 to 10 million reads (total ~97.7 million reads with detected primers and Q score of at least 9). Transcriptome reconstruction using IsoQuant initially identified 15,933 novel isoforms (1,701 genes) relative to NCBI RefSeq (GRCg7b). Using SQANTI3, we filtered transcript models with evidence of RT-switching, intrapriming, non-canonical splice sites, or limited sample support. These stringent filtration criteria retained 4,005 high-confidence novel isoforms (215 genes). After cross-validating against the extended GEGA annotation, 1,106 of 4,005 novel isoforms (27.6%) and 157 of 215 novel genes (73%) were identified to be already annotated. Recovering annotations that are independently annotated in GEGA provides strong validation for our discovery pipeline. Taken together, we have identified 2,899 isoforms and 58 genes that are novel relative to both NCBI RefSeq and GEGA annotations. We developed and validated a high-confidence chicken breast-muscle isoform annotation to enable isoform-resolved investigation of WB. Furthermore, our study provides an extended annotation for isoform-resolved functional characterization of known QTLs in livestock production traits and isoform dynamics across myogenic development in chicken breast muscle.