Modeling Species Specific Gene Expression Across Multiple Regions in the Brain
- 55 Downloads
Motivation: The question of what makes the human brain functionally different from that of other closely related primates, such as the chimpanzee, has both philosophical as well as practical implications. One of the challenges faced with such studies, however, is the small sample size available. Furthermore, expression values for multiple brain regions have an inherent structure that is generally ignored in published studies.
Results: We present a new statistical approach to identify genes with species specific expression, that (1) avoids multiple pairwise comparisons, which can be susceptible to small changes in expression as well as intransitivity, and (2) pools information across related data sets when available to produce more robust results, such as in the case of gene expression across multiple brain regions. We demonstrate through simulations that our model can much better identify human specific genes than the naive approach. Applications of the model to two previously published data sets, one microarray and one RNA-Seq, suggest a moderately large benefit from our model. We show that our approach produces more robust gene classifications across regions, and greatly reduces the number of human specific genes previously reported, which we show were primarily due to the noise in the underlying data.
KeywordsGene expression R code Posterior probabilities Markov random field RNA sequencing Akaike Bayes
We would like to thank Zhixiang Lin, for discussion of the Markov random field model and its applications.
Funding LD was supported by the National Library of Medicine Informatics training grant. HZ was supported in part by NIH R01 GM59507.
- 1.Akaike, H. (1973). Information theory and an extension of the maximum likelihood principle. In B. N. Petrov & F. Csaki (Eds.). Second International Symposium on Information Theory (pp. 267–281). Budapest: Akadémiai Kiado.Google Scholar
- 2.Anders, S., & Huber, W. (2010). Differential expression analysis for sequence count data. Genome Biology, 11(10).Google Scholar
- 4.Cáceres, M., Lachuer, J., Zapala, M. A., Redmond, J. C., Kudo, L., Geschwind, D. H., et al. (2003). Elevated gene expression levels distinguish human from non-human primate brains. Proceedings of the National Academy of Sciences of the United States of America, 100(22), 13030–13035. ISSN 0027-8424. https://doi.org/10.1073/pnas.2135499100 CrossRefGoogle Scholar
- 6.Dayton, C. M. (1998). Information criteria for the paired-comparisons problem. The American Statistician, 52(2), 144–151.Google Scholar
- 8.Delmar, P., Robin, S., Daudin, J., Delmar, P., Robin, S., & Daudin, J. (2005). VarMixt: Efficient variance modelling for the differential analysis of replicated gene expression data. ISSN 1367-4803. https://doi.org/10.1093/bioinformatics/bti023
- 10.Florence, J., Guillemette, M., Séverine, D., Isabelle, H., & Jean-Louis, F. (2007). A structural mixed model for variances in differential gene expression studies. p. 19. ISSN 0016-6723. https://doi.org/10.1017/S0016672307008646
- 12.Jeanmougin, M., de Reynies, A., Marisa, L., Paccard, C., Nuel, G., & Guedj, M. (2010). Should we abandon the t-test in the analysis of gene expression microarray data: A comparison of variance modeling strategies. Plos One, 5(9), e12336. https://doi.org/10.1371/journal.pone.0012336 CrossRefGoogle Scholar
- 16.Lin, Z., Li, M., Sestan, N., & Zhao, H. (2016). A Markov random field-based approach for joint estimation of differentially expressed genes in mouse transcriptome data. Statistical Applications in Genetics and Molecular Biology. https://doi.org/10.1515/sagmb-2015-0070. http://www.ncbi.nlm.nih.gov/pubmed/26926866
- 17.Lin, Z., Sanders, S. J., Li, M., Sestan, N., State, M. W., & Zhao, H. (2015). A Markov random field-based approach to characterizing human brain development using spatial-temporal transcriptome data. Annals of Applied Statistics, 9(1), 429–451. https://doi.org/10.1214/14-AOAS802. http://www.ncbi.nlm.nih.gov/pubmed/26877824