An mRNA expression atlas for the domestic chicken generated by the meta-analysis of RNA-seq datasets

  • Stephen Bush (Creator)
  • Lucy Freem (Creator)
  • Amanda MacCallum (Creator)
  • Chunlei Wu (Creator)
  • Cyrus Afrasiabi (Creator)
  • Jenny O'Dell (Creator)
  • Androniki Psifidi (Creator)
  • Mark Stevens (Creator)
  • Jacqueline Smith (Creator)
  • David Hume (Creator)



18 RNA-sequencing datasets, of chicken (Gallus gallus) bone marrow derived macrophages (BMDMs), both unstimulated and 24 hours after stimulation with lipopolysaccharide (LPS), are available from the European Nucleotide Archive under study accession PRJEB22373 ( These samples are derived from 3 male and 3 female broilers (Ross 308), and 3 female layers (CSF1R-MacApple transgenic NOVOgen Brown), all 8 weeks of age.
The dataset attached here contains expression level estimates per gene (as transcripts per million), generated after randomly down-sampling each of the 18 BMDM +/- LPS datasets to 10 million reads 100 times (using seqtk seeded with a random integer between 0 and 10,000). It may be seen that the random sampling used to down-size each dataset does not substantially alter the relative expression estimates of any two genes within any given sample, with equivalent expression profiles reconstructed for each of 100 random samples.


The chicken (Gallus gallus) is widely used as a model in developmental biology and is also an important livestock species. We have created an mRNA expression atlas for the domestic chicken spanning major tissue types and developmental stages, using a diverse range of publicly-archived RNA-seq datasets and new data derived from immune cells and tissues. Randomly down-sampling each dataset to a common depth and quantifying expression against a reference transcriptome using the probabilistic mRNA quantitation tool Kallisto ensures that each dataset explores comparable transcriptomic space, despite methodological differences between laboratories. The network analysis tool Miru was used to extract robust clusters of co-expressed genes from the atlas. Many clusters were strongly tissue or cell-type restricted, including a set of macrophage-specific transcripts. Other co-expression clusters were clearly associated with processes, including the cell division cycle and the formation and function of motile cilia. In many cases, the clusters contain transcription factors that have previously been implicated in their regulation. For example, the macrophage cluster includes BATF3, CEBPB, IRF1, NFE2L2, NRR1H3, SPI1, STAT1, and TFEC, each of which is also macrophage-enriched in mammals. Likewise, the cell cycle cluster contains FOXM1 and the motile cilia cluster contains FOXJ1. The atlas therefore provides a resource for the annotation of genes that currently have only a locus ID based upon the principle of ‘guilt by association’. The approach we have developed can be extended with new datasets from novel tissues, and is applicable to any species.

Data Citation

Bush, Stephen J.; Freem, Lucy; MacCallum, Amanda; O'Dell, Jenny; Wu, Chunlei; Afrasiabi, Cyrus; Psifidi, Androniki; Stevens, Mark P.; Smith, Jacqueline; Hume, David A.. (2017). An mRNA expression atlas for the domestic chicken generated by the meta-analysis of RNA-seq datasets, [dataset]. The Roslin Institute, University of Edinburgh.
Date made available6 Sept 2017
PublisherEdinburgh DataShare

Cite this