as being publicly available on the websites, which usually is very well received by journals and reviewers on future grant applications. Many of the current public repositories will not accept files of the size of zoomable histopathology images and limited user uploads. This, together with payment to commercial repositories is a specific problem for very image intensive histopathology and cell biology data.
General Information on Inbred Laboratory Mice
Mouse Genome Informatics (http://www.informatics.jax.org)
The Mouse Genome Informatics (MGI) international database resource is the most comprehensive collection of information on mouse biology available today. It is the first place one should go to when searching for information on the mouse or even just searching for a mouse. MGI is an integrated collection of databases providing access to comprehensive information on genetics, genomics, and biology of the laboratory mouse. Data in MGI are obtained by direct expert curation of published literature and importation of large‐scale datasets from other data resources. The Mouse Genome Database (MGD) [4] is the core resource for general mouse information in MGI. MGD contains information on mouse genes, strains, phenotypes, mutant alleles, sequences/single nucleotide polymorphisms (SNP), vertebrate homology, and mouse models of human disease. MGD also is the source for official gene, allele, and mouse strain nomenclature using International Committee on Standardized Nomenclature for Mice standards (see Chapter 3 for details on nomenclature). The GXD [5] contains information on gene expression in the developing mouse including immunohistochemistry, immunofluorescence, and in situ hybridization images. The CrePortal contains data on cre‐recombinase expressing mice with a large library of images [6]. The Gene Ontology (GO) project provides functional annotations for mouse gene products [7–11]. The Mouse Models of Human Cancer database (MMHC) contains information on inbred mouse strains, genetically engineered mouse (GEM) models, [12–15] and patient derived xenograft (PDX) models of human cancer [16]. This database is particularly useful for pathologists and is described in detail below. Databases in the MGI consortium that contain images are the MGD, the MMHC, and the GXD.
MouseMine (http://www.mousemine.org/mousemine/begin.do)
MouseMine [17] is a data warehouse resulting from a collaboration between the Intermine project at the University of Cambridge and The Jackson Laboratory MGI group. MouseMine contains a subset of data from MGI, focused on genomics, strain, and allele annotations curated by MGI, such as GO [7–11] and Mutant Phenotype using Mouse Phenotype Ontology [18]. Specifically, it is focused on comparative human and mouse data and contains a great deal of information annotated to standard human ontologies such as the Disease Ontology [19] and the Human Phenotype Ontology [20]. It is cross‐referenced to Online Mendelian Inheritance in Man (OMIM) [21] (see below) and other external resources. A full list of the data in MouseMine can be found at http://www.mousemine.org/mousemine/dataCategories.do. One of the strengths of MouseMine is that it is possible to carry out very sophisticated queries, to integrate the results of MouseMine queries with user data, and to combine the outputs of multiple queries. It also offers web services over an API to provide programmatic access to MGI data.
Monarch (https://monarchinitiative.org)
Monarch [22] aims to provide a formally integrated resource pulling together data from model organisms such as the mouse, and from humans, through alignment of genotypes, variants, diseases, and phenotypes. The focus is on using formally defined phenotypes to support identification of animal models for human disease. The database pulls data from wide range of sources (Figure 2.1) and provides powerful methods for data exploration. Data can be accessed through an API and RESTful web services.
Alliance of Genome Resources (https://www.alliancegenome.org)
The Alliance of Genome Resources is a consortium of six model organism databases (MODs) and the (GO) consortium. The purpose of the alliance is to provide a central data resource that provides a common user interface to query genomic data and integrates the data from different organisms allowing this data to be viewed in a consistent manner. The Alliance also provides links back to the primary data in many cases. The ability to compare common data across multiple model organisms in an integrated manner is a very useful and powerful tool [23]. The MGD is one of the founding members of the Alliance.
Figure 2.1 Data integration strategy for Monarch. Data, mainly “genotype‐to‐phenotype,” are gathered from a wide range of databases and parsed so that they can be organized into an integrated knowledge graph based on the 35 underlying ontologies used by Monarch describing for example genotype–phenotype relations, diseases, and gene properties. The knowledge graph can then be browsed, searched, or semantic similarity between entities calculated to look for related diseases or mutant strains, for example, where the relationship had previously not been noticed. The databases integrated by Monarch can be found on https://monarchinitiative.org/about/data‐sources.
Source: Image reproduced with permission from Monarch.
Mouse Models of Human Cancer Database (MMHC, Formerly Mouse Tumor Biology Database) (http://tumor.informatics.jax.org/mtbwi/index.do)
The Mouse Models of Human Cancer Database is one of the core databases included in the MGI consortium. Unlike Pathbase (see below), which covers all types of lesions diagnosed in laboratory mice, this website focuses on cancer and hyperplastic lesions in mice (Figure 2.2). Much of the database is built around data and the literature such that the images are supplemental to the enormous amount of data on strain specific lesions, genes known to cause specific types of cancer, and how this information relates to human cancer [12,24–29]. A recent change has been to integrate photomicrographs of human cancers into the database for more direct anatomic comparisons. Also, MMHC contains datasets and webpages focused on specific topics, such as skin cancers [15], lymphomas [13], and a list of antibodies tested in mice with conditions tested under and results [30]. Whole slide scanned images of the lymphomas with immunohistochemistry demonstrate how to subtype the lesions can be accessed in this database [13]. The lymphoma digital slide (whole slide images) collection originates from the NCI Mouse Models of Cancer Consortium workshop on Hematopoietic Tumors (http://tumor.informatics.jax.org/mtbwi/lymphomaPathology.jsp) [31, 32]. Lung cancers and hyperplasias were whole‐slide scanned from a large‐scale aging program looking at lesions in 31 inbred mouse strains [33, 34]. Current projects include developing a training module to aid pathologists in the interpretation and grading of non‐melanoma skin cancers in mice caused by UV light, two‐stage chemical carcinogenesis, papillomavirus, or those that arise in production colonies spontaneously. As PDX (cancers resected from patients and transplanted into immunodeficient mice) have become a major research tool, these too are included in this database [16].
The histopathology images included in MMHC are curated from the primary literature, when given permission by the publisher, and from direct submissions from researchers and large‐scale projects.
As