Search

Background Differential isoform usage is an important driver of inter-individual phenotypic diversity and is linked to various diseases and traits. However, accurately detecting the differential usage of different gene transcripts between groups can be difficult, in particular in less well annotated genomes where the spectrum of transcript isoforms is largely unknown. Results We investigated whether machine learning approaches can detect differential isoform usage based purely on the distribution of reads across a gene region. We illustrate that gradient boosting and elastic net approaches can successfully identify large numbers of genes showing potential differential isoform usage between Europeans and Africans, that are enriched among relevant biological pathways and significantly overlap those identified by previous approaches. We demonstrate that diversity at the 3′ and 5′ ends of genes are primary drivers of these differences between populations. Conclusion Machine learning methods can effectively detect differential isoform usage from read fraction data, and can provide novel insights into the biological differences between groups.

Additional file 3. Figures S1 to S5 and Tables S1 to S2.

This study highlights dangers in over-interpreting fine-mapping results. Chundru et al. show that genotype imputation accuracy has a large impact on fine-mapping accuracy. They used DNA methylation at CpG-sites with a variant..., Genetic variants disrupting DNA methylation at CpG dinucleotides (CpG-SNP) provide a set of known causal variants to serve as models to test fine-mapping methodology. We use 1716 CpG-SNPs to test three fine-mapping approaches (Bayesian imputation-based association mapping, Bayesian sparse linear mixed model, and the J-test), assessing the impact of imputation errors and the choice of reference panel by using both whole-genome sequence (WGS), and genotype array data on the same individuals (n = 1166). The choice of imputation reference panel had a strong effect on imputation accuracy, with the 1000 Genomes Project Phase 3 (1000G) reference panel (n = 2504 from 26 populations) giving a mean nonreference discordance rate between imputed and sequenced genotypes of 3.2% compared to 1.6% when using the Haplotype Reference Consortium (HRC) reference panel (n = 32,470 Europeans). These imputation errors had an impact on whether the CpG-SNP was included in the 95% credible set, with a difference of ∼23% and ∼7% between the WGS and the 1000G and HRC imputed datasets, respectively. All of the fine-mapping methods failed to reach the expected 95% coverage of the CpG-SNP. This is attributed to secondary cis genetic effects that are unable to be statistically separated from the CpG-SNP, and through a masking mechanism where the effect of the methylation disrupting allele at the CpG-SNP is hidden by the effect of a nearby SNP that has strong linkage disequilibrium with the CpG-SNP. The reduced accuracy in fine-mapping a known causal variant in a low-level biological trait with imputed genetic data has implications for the study of higher-order complex traits and disease.

Goniodysgenesis is a developmental abnormality of the anterior chamber of the eye. It is generally considered to be congenital in dogs (Canis lupus familiaris), and has been associated with glaucoma and blindness. Goniodysgenesis and early-onset glaucoma initially emerged in Border Collies in Australia in the late 1990s and have subsequently been found in this breed in Europe and the USA. The objective of the present study was to determine the genetic basis of goniodysgenesis in Border Collies. Clinical diagnosis was based on results of examinations by veterinary ophthalmologists of affected and unaffected dogs from eleven different countries. Genotyping using the Illumina high density canine single nucleotide variant genotyping chip was used to identify a candidate genetic region. There was a highly significant peak of association over chromosome 17, with a p-value of 2 × 10−13. Expression profiles and evolutionary conservation of candidate genes were assessed using public databases. Whole genome sequences of three dogs with glaucoma, three severely affected by goniodysgenesis and three unaffected dogs identified a missense variant in the olfactomedin like 3 (OLFML3) gene in all six affected animals. This was homozygous for the risk allele in all nine cases with glaucoma and 12 of 14 other severely affected animals. Of 67 reportedly unaffected animals, only one was homozygous for this variant (offspring of parents both with goniodysgenesis who were also homozygous for the variant). Analysis of pedigree information was consistent with an autosomal recessive mode of inheritance for severe goniodysgenesis (potentially leading to glaucoma) in this breed. The identification of a candidate genetic region and putative causative variant will aid breeders to reduce the frequency of goniodysgenesis and the risk of glaucoma in the Border Collie population.

X chromosomes are unusual in many regards, not least of which is their nonrandom gene content. The causes of this bias are commonly discussed in the context of sexual antagonism and the avoidance of activity in the male germline. Here, we examine the notion that, at least in some taxa, functionally biased gene content may more profoundly be shaped by limits imposed on gene expression owing to haploid expression of the X chromosome. Notably, if the X, as in primates, is transcribed at rates comparable to the ancestral rate (per promoter) prior to the X chromosome formation, then the X is not a tolerable environment for genes with very high maximal net levels of expression, owing to transcriptional traffic jams. We test this hypothesis using The Encyclopedia of DNA Elements (ENCODE) and data from the Functional Annotation of the Mammalian Genome (FANTOM5) project. As predicted, the maximal expression of human X-linked genes is much lower than that of genes on autosomes: on average, maximal expression is three times lower on the X chromosome than on autosomes. Similarly, autosome-to-X retroposition events are associated with lower maximal expression of retrogenes on the X than seen for X-to-autosome retrogenes on autosomes. Also as expected, X-linked genes have a lesser degree of increase in gene expression than autosomal ones (compared to the human/Chimpanzee common ancestor) if highly expressed, but not if lowly expressed. The traffic jam model also explains the known lower breadth of expression for genes on the X (and the Z of birds), as genes with broad expression are, on average, those with high maximal expression. As then further predicted, highly expressed tissue-specific genes are also rare on the X and broadly expressed genes on the X tend to be lowly expressed, both indicating that the trend is shaped by the maximal expression level not the breadth of expression per se. Importantly, a limit to the maximal expression level explains biased tissue of expression profiles of X-linked genes. Tissues whose tissue-specific genes are very highly expressed (e.g., secretory tissues, tissues abundant in structural proteins) are also tissues in which gene expression is relatively rare on the X chromosome. These trends cannot be fully accounted for in terms of alternative models of biased expression. In conclusion, the notion that it is hard for genes on the Therian X to be highly expressed, owing to transcriptional traffic jams, provides a simple yet robustly supported rationale of many peculiar features of X’s gene content, gene expression, and evolution., Laurence Hurst, Lukasz Huminiecki, and the FANTOM5 consortium propose a new explanation for the peculiar expression properties of genes on the human X chromosome, based on the premise that very high expression levels cannot be achieved on a haploid-expressed chromosome., Author Summary Genes located on the human X chromosome are not a random mix of genes: they tend to be expressed in relatively few tissues or are specific for a particular set of tissues, e.g., brain regions. Prior attempts to explain this skewed gene content have hypothesized that the X chromosome might be peculiar because it has to balance mutations that are advantageous to one sex but deleterious to the other, or because it has to shut down during the process of sperm manufacture in males. Here we suggest and test a third possible explanation: that genes on the X chromosome are limited in their transcription levels and thus tend to be genes that are lowly or specifically expressed. We consider the suggestion that since these genes can only be expressed from one chromosome, as males only have one X, the ability to express a gene at very high rates is limited owing to potential transcriptional traffic jams. As predicted, we find that human X-located genes have maximal expression rates far below that of genes residing on autosomes. When we look at genes that have moved onto or off the X chromosome during recent evolution, we find the maximal expression is higher when not on the X chromosome. We also find that X-located genes that are relatively highly expressed are not able to increase their expression level further. Our model explains both the enrichment for tissue specificity and the paucity of certain tissues with X-located genes. Genes underrepresented on the X are either expressed in many tissues—such genes tend to have high maximal expression—or are from tissues that require a lot of transcription (e.g., fast secreting tissues like the liver). Just as many of the findings cannot be explained by the two earlier models, neither can the traffic jam model explain all the peculiar features of the genes found on the X chromosome. Indeed, we find evidence of a reproduction-related bias in X-located genes, even after allowing for the traffic jam problem.