Developmental gene expression differences between humans and mammalian models

Abstract

14Identifying the molecular programs underlying human organ development and how they differ from 15 those in model species will advance our understanding of human health and disease. 16 Developmental gene expression profiles provide a window into the genes underlying organ 17 development as well as a direct means to compare them across species. We use a transcriptomic 18 resource for mammalian organ development to characterize the temporal profiles of human genes 19 associated with distinct disease classes and to determine, for each human gene, the similarity of its 20 spatiotemporal expression with its orthologs in rhesus macaque, mouse, rat and rabbit. We find 21 that half of human genes differ from their mouse orthologs in their temporal trajectories. These 22 include more than 200 disease genes associated with brain, heart and liver disease, for which mouse 23 models should undergo extra scrutiny. We provide a new resource that evaluates for every human 24 gene its suitability to be modeled in different mammalian species. 25 26 27 28 42 rhesus macaques) are routinely used as models of both normal human development and human 43 disease because it is generally assumed that the genes and regulatory networks underlying 44 development are largely conserved across these species. While this is generally true, there are also 45 critical differences between species during development, which underlie the large diversity of 46 3 mammalian organ phenotypes [1-6,8]. Identifying the commonalities and differences between the 47 genetic programs underlying organ development in different mammalian species is therefore 48 paramount for assessing the translatability of knowledge obtained from mammalian models to 49 understand human health and disease. Critically, gene expression profiles can be directly compared 50 between species, especially when they are derived from matching cells/organs and developmental 51 stages. Gene expression therefore offers a direct means to evaluate similarities and differences 52 between species in organ developmental programs. While the relationship between gene 53 expression and phenotypes is not linear, identifying when and where gene expression differs 54 between humans and other species will help identify the conditions (i.e., developmental stages, 55 organs, genes) under which model species may not be well suited to model human development 56 and disease. 57 58Here, we take advantage of a developmental gene expression resource [13], which densely covers 59 the development of seven major organs in humans and other mammals, to characterize the 60 spatiotemporal profiles of human disease genes and gain new insights into the symptomatology of 61 diseases. We also determine for each human gene (including disease-associated genes) the 62 similarity of its spatiotemporal expression with that of its orthologs in mouse, rat, rabbit and rhesus 63 macaque. Our analyses and datasets therefore provide a new resource for assessing the suitability 64 of different mammalian species to model the action ...

Mendelian gene identification through mouse embryo viability screening

“…To assess whether the organ development trajectories for these genes differ substantially between mouse and human, we investigated the similarity of spatiotemporal gene expression profiles for the two species. We found that 78 and 82% of the entire set of genes under study showed the same trajectory for cerebellum and brain respectively, with no significant differences observed between WoL, and in concordance with what was observed for the entire set of genes with data available 25 ( Sup Fig 2 ). Similarities in gene expression will not always imply conserved phenotypes between mouse and human, but they can serve as a proxy for how translatable to human disease the findings for these genes are.…”

“…Human gene expression (RPKM) across development for brain, cerebellum, heart, kidney, liver, ovary and testis was obtained from Cardoso-Moreira et al 23 https://apps.kaessmannlab.org/evodevoapp/ [Downloaded 10.08 .21] Data on comparison of temporal trajectories between human genes and their orthologs in mouse for brain and cerebellum was obtained from Cardoso-Moreira et al 25 .…”

Cellular development and evolution of the mammalian cerebellum

“…9d). In line with studies linking phenotypic severity to expression pleiotropy 19,41 , we found that the dynamic genes shared between the two neuron types are under stronger constraint than genes dynamic in one cell type only (Fig. 3e, Extended Data Fig.…”

“…We used different metrics of gene essentiality: (i) LOEUF score (loss-of-function observed/expected upper bound fraction) 39 from the Genome Aggregation Database (gnomAD), which ranks genes along a continuous spectrum of tolerance to loss-of-function variation based on human exome and genome sequencing data (in vivo); (ii) a measure of in vivo essentiality that combines various essentiality scores (RVIS, pLI, Phi, missense Z-score, LoFtool and s het ) to rank genes based on human exome and genome sequencing data 38 ; (iii) a measure of in vitro essentiality that combines viability scores from CRISPR-Cas9 screens on human cell lines 38 . We obtained the human inherited disease gene list from the manually curated Human Gene Mutation Database (HGMD, PRO 17.1) 42 , and only used genes with disease-causing mutations as described previously 41 . We used the genes that based on the mapping to the Unified Medical Language System (UMLS) were linked to the high level disease types 'Nervous system' or 'Psychiatric', and split these genes into two groups based on whether a gene was also linked to the high level disease type 'Developmental' (developmental n=373; other n=200).…”

Antibodies, Nanobodies, or Aptamers—Which Is Best for Deciphering the Proteomes of Non-Model Species?

“…Proteins known to be expressed in humans were also found to be enriched in frogs, but surprisingly not in mice or pigs [35]. It is unsurprising then that frogs may find their way into more translatable research in the future [36], especially given that half of human genes differ from their mouse orthologs in different developmental trajectories, including more than 200 disease genes associated with brain, heart, and liver diseases [37]. These differences can impact the proteomes, and therefore phenotypes, between humans and mice.…”