The Essentiality Status of Mouse Duplicate Gene Pairs Correlates with Developmental Co-Expression Patterns

Interspecific comparison of gene expression profiles using machine learning

Interspecific gene comparisons are the keystones for many areas of biological research and are especially important for the translation of knowledge from model organisms to economically important species. Currently they are hampered by the low resolution of methods based on sequence analysis and by the complex evolutionary history of eukaryotic genes. This is especially critical for plants, whose genomes are shaped by multiple whole genome duplications and subsequent gene loss. This requires the development of new methods for comparing the functions of genes in different species. Here, we report ISEEML (Interspecific Similarity of Expression Evaluated using Machine Learning)-a novel machine learning-based algorithm for interspecific gene classification. In contrast to previous studies focused on sequence similarity, our algorithm focuses on functional similarity inferred from the comparison of gene expression profiles. We propose novel metrics for expression pattern similarity-expression score (ES)-that is suitable for species with differing morphologies. As a proof of concept, we compare detailed transcriptome maps of Arabidopsis thaliana, the model species, Zea mays (maize) and Fagopyrum esculentum (common buckwheat), which are species that represent distant clades within flowering plants. The classifier resulted in an AUC of 0.91; under the ES threshold of 0.5, the specificity was 94%, and sensitivity was 72%.

Mendelian gene identification through mouse embryo viability screening

BACKGROUND

The diagnostic rate of Mendelian disorders in sequencing studies continues to increase, along with the pace of novel disease gene discovery. However, variant interpretation in novel genes not currently associated with disease is particularly challenging and strategies combining gene functional evidence with approaches that evaluate the phenotypic similarities between patients and model organisms have proven successful. A full spectrum of intolerance to loss-of-function variation has been previously described, providing evidence that gene essentiality should not be considered as a simple and fixed binary property.

METHODS

Here we further dissected this spectrum by assessing the embryonic stage at which homozygous loss-of-function results in lethality in mice from the International Mouse Phenotyping Consortium, classifying the set of lethal genes into one of three windows of lethality: early, mid, or late gestation lethal. We studied the correlation between these windows of lethality and various gene features including expression across development, paralogy and constraint metrics together with human disease phenotypes. We explored a gene similarity approach for novel gene discovery and investigated unsolved cases from the 100,000 Genomes Project.

RESULTS

We found that genes in the early gestation lethal category have distinct characteristics and are enriched for genes linked with recessive forms of inherited metabolic disease. We identified several genes sharing multiple features with known biallelic forms of inborn errors of the metabolism and found signs of enrichment of biallelic predicted pathogenic variants among early gestation lethal genes in patients recruited under this disease category. We highlight two novel gene candidates with phenotypic overlap between the patients and the mouse knockouts.

CONCLUSIONS

Information on the developmental period at which embryonic lethality occurs in the knockout mouse may be used for novel disease gene discovery that helps to prioritise variants in unsolved rare disease cases.

Functional compensation of mouse duplicates by their paralogs expressed in the same tissues

Analyses in a number of organisms have shown that duplicated genes are less likely to be essential than singletons. This implies that genes can often compensate for the loss of their paralogs. However, it is unclear why the loss of some duplicates can be compensated by their paralogs, whereas the loss of other duplicates cannot. Surprisingly, initial analyses in mice did not detect differences in the essentiality of duplicates and singletons. Only subsequent analyses, using larger gene knockout datasets and controlling for a number of confounding factors, did detect significant differences. Previous studies have not taken into account the tissues in which duplicates are expressed. We hypothesized that in complex organisms, in order for a gene's loss to be compensated by one or more of its paralogs, such paralogs need to be expressed in at least the same set of tissues as the lost gene. To test our hypothesis, we classified mouse duplicates into two categories based on the expression patterns of their paralogs: "compensable duplicates" (those with paralogs expressed in all the tissues in which the gene is expressed) and "non-compensable duplicates" (those whose paralogs are not expressed in all the tissues where the gene is expressed). In agreement with our hypothesis, the essentiality of non-compensable duplicates is similar to that of singletons, whereas compensable duplicates exhibit a substantially lower essentiality. Our results imply that duplicates can often compensate for the loss of their paralogs, but only if they are expressed in the same tissues. Indeed, the compensation ability is more dependent on expression patterns than on protein sequence similarity. The existence of these two kinds of duplicates with different essentialities, which has been overlooked by prior studies, may have hindered the detection of differences between singletons and duplicates.

On the relation of gene essentiality to intron structure: a computational and deep learning approach

Essential genes have been studied by copy number variants and deletions, both associated with introns. The premise of our work is that introns of essential genes have distinct characteristic properties. We provide support for this by training a deep learning model and demonstrating that introns alone can be used to classify essentiality. The model, limited to first introns, performs at an increased level, implicating first introns in essentiality. We identify unique properties of introns of essential genes, finding that their structure protects against deletion and intron-loss events, especially centered on the first intron. We show that GC density is increased in the first introns of essential genes, allowing for increased enhancer activity, protection against deletions, and improved splice site recognition. We find that first introns of essential genes are of remarkably smaller size than their nonessential counterparts, and to protect against common 3' end deletion events, essential genes carry an increased number of (smaller) introns. To demonstrate the importance of the seven features we identified, we train a feature-based model using only these features and achieve high performance.

Dedicated transcriptomics combined with power analysis lead to functional understanding of genes with weak phenotypic changes in knockout lines

Systematic knockout studies in mice have shown that a large fraction of the gene replacements show no lethal or other overt phenotypes. This has led to the development of more refined analysis schemes, including physiological, behavioral, developmental and cytological tests. However, transcriptomic analyses have not yet been systematically evaluated for non-lethal knockouts. We conducted a power analysis to determine the experimental conditions under which even small changes in transcript levels can be reliably traced. We have applied this to two gene disruption lines of genes for which no function was known so far. Dedicated phenotyping tests informed by the tissues and stages of highest expression of the two genes show small effects on the tested phenotypes. For the transcriptome analysis of these stages and tissues, we used a prior power analysis to determine the number of biological replicates and the sequencing depth. We find that under these conditions, the knockouts have a significant impact on the transcriptional networks, with thousands of genes showing small transcriptional changes. GO analysis suggests that A930004D18Rik is involved in developmental processes through contributing to protein complexes, and A830005F24Rik in extracellular matrix functions. Subsampling analysis of the data reveals that the increase in the number of biological replicates was more important that increasing the sequencing depth to arrive at these results. Hence, our proof-of-principle experiment suggests that transcriptomic analysis is indeed an option to study gene functions of genes with weak or no traceable phenotypic effects and it provides the boundary conditions under which this is possible.

Knockout mice benefit the understanding of gene functions in mammals. However, it has proven difficult for many genes to identify clear phenotypes, related due to lack of sufficient assays. As Lewis Wolpert put it in a famous quote “But did you take them to the opera?”, thus metaphorically alluding to the need to extend phenotyping efforts. This insight led to the establishment of phenotyping pipelines that are nowadays routinely used to characterize knock-out lines. However, transcriptomic approaches based on RNA-Seq have been much less explored for such deep-level studies. We conducted here both, a theoretical power analysis and practical RNA-Seq experiments on two knockout lines with small phenotypic effects to investigate the parameters including sample size, sequencing depth, fold change, and dispersion. Our dedicated RNA-Seq studies discovered thousands of genes with small transcriptional changes and enriched in specific functions in both knockout lines. We find that it is more important to increase the number of samples than to increase the sequencing depth. Our work shows that a deep RNA-Seq study on knockouts is powerful for understanding gene functions in cases of weak phenotypic effects, and provides a guideline for the experimental design of such studies.