Human and mouse platelet transcriptomes and proteomes for phenotyping 3474 genes with hemostatic and platelet traits

Abstract

The hemostatic process relies on platelet and coagulation activation, with additional roles of red blood cells and the vessel wall. By systematic screening of databases for gene-linked information on hemostasis, we collected phenotypic profiles of 3474 orthologous human and mouse genes regarding bleeding, arterial thrombosis, thrombophilia, platelet traits, coagulation, and erythrocytes. Comparisons showed that defects in 252 mouse genes led to increased bleeding combined with platelet dysfunction or thrombocytopenia, in addition to 150 human orthologs that are registered for familial bleeding disorders, based on panel sequencing. Additionally, 139 mouse genes contributed to arterial thrombosis without bleeding phenotype. To further investigate the role of platelets in hemostasis, we integrated multiple genome-wide RNA-sequencing transcriptomes and proteomes from healthy subjects and C57BL/6 mice. This provided reference levels for 54 790 (54 247) transcripts and 6379 (4563) proteins in human (mouse) platelets. Orthologous transcripts in human and mouse platelets correlated with R=0.75, whereas orthologous platelet proteins correlated with R=0.87. Comparison with the phenotypic analysis revealed the following: (i) overall high qualitative similarity of human and mouse platelets regarding composition and function; (ii) presence of transcripts in platelets for most of the 3474 phenotyped genes; (iii) preponderance of syndromic platelet-expressed genes; and (iv) 20-40% overlap with genes from genome-wide association studies. For 42 mouse genes, among which receptors, signaling proteins, and transcription regulators (ASXL1, ERG, GATA2, MEIS1, NFE2, and TAL1), we confirmed novel links with human platelet function or count. This interspecies comparison can serve as a valuable resource for researchers and clinicians studying the genetics of blood-borne hemostasis and thrombosis.

Graphical abstract

Figure 1.

Human and mouse genes linked to hemostasis or platelet-related phenotypes. (A) Numbers are shown of phenotyped orthologous human and mouse genes with scores for bleeding, thrombophilia (arterial thrombosis for mice), platelet function/morphology/count, coagulation (including fibrinolysis), and erythrocyte characteristics. (B) Classification scheme used for protein localization and function (classes C01-C22). (C) Human genes linked to bleeding and platelet traits per class. Colored bars indicate reference genes linked to platelet function (orange), thrombocytopenia (black) or both (brown). Gray boxes represent novel genes (names indicated); for phenotypic descriptions and data sources, see supplemental Table 13.

Figure 2.

Distribution profiles of orthologous transcripts in human and mouse platelets per function class. Quantified transcripts in human platelets (n = 54 790) and mouse platelets (n = 54 247) were assigned to 23 protein function classes, RNA genes, or pseudogenes. Histograms are given per class (100%) of mean fpkm/rpkm values (log2+1) for human platelets (A) and mouse platelets (B). (C) Percentage fractions of relevant transcripts (>0.03) per class (total 100%). (D) Correlation of relevant transcripts per class between human and mouse platelets. Plt, platelet.

Figure 3.

Distribution profiles of orthologous platelet proteins in human and mouse platelets per function class. Quantified proteins in human platelets (n = 6379) and mouse platelets (n = 4684) were assigned to 23 protein function classes. Histograms are given per class (100%) of mean copy numbers for human platelets (A) and mouse platelets (B). (C) Percentage fractions of identified proteins per class (total 100%). (D) Correlation of identified proteins per class between human and mouse platelets. Plt, platelet.

Figure 4.

Similarity of orthologous transcripts and proteins in human and mouse platelets. (A) Plot of transcript values [log₂(tpm + 1); tpm normalized; means of 5 transcriptomes] comparing human and mouse platelets. (B) Plot of protein copy numbers (calibrated; means of 5 proteomes) comparing human and mouse platelets. Shown are results of regression analysis with indicated number of gene products. (C) Partial overlap (in black) of orthologous transcripts in human and mouse platelets per function class (C01-C23; RNA genes; pseudogenes). The relevant transcript level was set at 0.03. (D) Partial overlap (in black) of detected proteins in human and mouse platelets per function class. Not identified means absent in proteome but present in transcriptome. Plt, platelets.

Figure 5.

Phenotype profiles of nonsyndromic genes linked to bleeding in combination with platelet dysfunction in mouse. Syndromic genes from supplemental File 1 were filtered for bleeding plus platelet dysfunction in mouse. Indicated in blue type are orthologous human genes from the reference list (known link to hemostasis). Phenotype scoring for humans (h) and mice (m) included the following: 1, bleeding; 2, thrombophilia/arterial thrombosis tendency; 3, platelet function; 4, platelet morphology and development; 5, platelet count; 6, coagulation dysfunction; 7, erythrocyte characteristics; 8, platelet traits from human GWAS (supplemental Figure 1). Indicated are also mean levels of platelet and megakaryocyte transcripts and platelet protein copy numbers. Extended phenotype descriptions per gene are given in supplemental File 1. Coagul, coagulation; MGK, megakaryocyte; P-count, platelet count; P-morph, platelet morphology; P-mRNA, platelet mRNA.

Figure 6.

Phenotype profiles of syndromic genes linked to bleeding in combination with platelet dysfunction in mouse. Syndromic genes from supplemental File 1 were filtered for bleeding plus platelet dysfunction in mouse. Phenotype scoring for human (h) and mouse (m) included was as described in Figure 5.

Figure 7.

Phenotype profiles of genes linked to bleeding in combination with thrombocytopenia or coagulation defects in mouse. Filtering of supplemental File 1 was for nonsyndromic (A) or syndromic (B) thrombocytopenia or for nonsyndromic (C) or syndromic (D) coagulation defects. Reference human genes are indicated in blue type. Phenotype scoring was as indicated in Figure 5.