Fully haplotyped genome assemblies of healthy individuals reveal variability in 5'ss strength and support by splicing...

This work presents a comprehensive investigation of sequence variations at human 5' splice sites (5'ss), exploring their impact on 5'ss strength and predicted splicing regulatory protein (SRP) binding. Leveraging 44 high-quality genomes, with fully haplotyped assemblies, we were able to fully assess homozygous and heterozygous sequence variations around and within 5'ss. Variations showed differing tolerance levels in protein-coding and non-coding transcripts. Around half of 5'ss variations did not alter 5'ss strength (measured by the HBond score, that estimates binding of the 11 nucleotides at the free U1 spliceosomal RNA 5' end). Heterozygous variations resulted in stronger 5'ss strength reductions and less compensatory effects of multiple sequence variations at the same 5'ss, compared to homozygous variations. Additionally, we observed a slight balance between changes in 5'ss strength and predicted SRP binding. Strong 5'ss (HBond score >18.8) showed strength variations in both directions, as theoretically expected for random distributions, whereas weaker 5'ss consistently showed lower strength reductions than expected or achievable. Variations at 5'ss of essential genes were less frequent than in other genes and showed a higher amount of variations that did not alter 5'ss strength. Acceptable changes in predicted SRP binding sites were highly dependent on their respective 5'ss strength.