Experienced morphologists and physicians use HREM image data to systematically identify phenotypes in both mutant and wild-type embryos.

A mouse embryo develops rapidly between E14 and E15, and its precise morphology can change dramatically within this period. In addition, homozygous mutant embryos are frequently delayed in development. To avoid scoring phenotypes that are, in fact, normal features of development at a given time point, each embryo must be staged precisely before phenotypes are identified.


  • An external 3D volume rendered model of the embryo is produced using software such as Osirix or Amira. For speed of modelling, the dataset can be subsampled to 25% without any appreciable loss of resolution.
  • If there are no malformations of the forelimb the embryo is staged using a novel technique based solely on the external forelimb appearance (Geyer et al, 2017).
    • This method provides additional stage granularity, splitting the three traditional Theiler Stages TS21 – TS23 into six distinct stages: 21, 22-, 22, 22+, 23- and 23, based on the extent of inter-digital webbing (see Image 1).
  • If the forelimb is malformed, the embryo is staged using the crown-rump length, lower jaw length and/or tibia length, which were found to correlate with the forelimb appearance. (Geyer et al, 2017).
Image shows six distinct stages based on the extent of inter-digital webbing.

Gross morphology

The external embryo 3D model is used to identify gross morphological defects. These include surface abnormalities, as well as defects of the jaw, limbs and pinna.

Phenotype identification

Detailed phenotyping is based on 2D HREM section images, which are viewed in Osirix. This removes the need for time-consuming 3D modelling of the full resolution data HREM data.

Where an embryo deviates from the expected morphology at the developmental stage determined during the staging procedure, a phenotype is assigned following the Mammalian Phenotype ontology.

  • The majority of phenotypes are identified by examining the original 2D axial section images.
  • The dataset is then re-sectioned in the sagittal direction, to allow better visualisation of possible phenotypes in the urethra, anus, diaphragm and pericardium.
  • Finally, the dataset is re-sectioned in the coronal direction to identify the remaining phenotypes, which include those related to the semilunar valves and the number of ribs.
  • Phenotypes for each embryo are scored independently by at least two researchers. The results are then discussed and combined into one phenotype set.

The detailed order in which phenotypes are identified is described in Weninger et al, 2014.

Phenotype annotation

  • Each phenotype is identified as a unique point in 3D space, and marked with a bounding box (using a bespoke plugin for Osirix) in the 2D section of the HREM image for all 3 orthogonal planes.
  • Additional information about the phenotype or image is added as a comment.

Individual contributions

  • WJ Weninger – co-ordination; cross-checking wild-types and mutants
  • SH Geyer – staging and gross morphology
  • D Szumska – phenotype scoring and annotation for wild-types
  • L Reissig – phenotype scoring and annotation for mutants
  • J Rose – cross-checking wild-types


  • Geyer SH, Reissig L, Rose J, Wilson R, Prin F, Szumska D, Ramirez-Solis R, Tudor C, White J, Mohun TJ, Weninger WJ (2017)
    A staging system for correct phenotype interpretation of mouse embryos harvested on embryonic day 14 (E14.5).
    Journal of Anatomy doi: 10.1111/joa.12590 Full article on Wiley Online Library
  • Weninger WJ, Geyer SH, Martineau A, Galli A, Adams DJ, Wilson R and Mohun TJ (2014)
    Phenotyping structural abnormalities in mouse embryos using high-resolution episcopic microscopy.
    Disease Models & Mechanisms 7 , 1143-1152 PubMed abstract