RT Journal Article
SR Electronic
T1 Not by systems alone: identifying functional outliers in rare disease pedigrees
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 128439
DO 10.1101/128439
A1 Sara Ballouz
A1 Max Dörfel
A1 Jonathan Crain
A1 Megan Crow
A1 Laurence Faivre
A1 Catherine E. Keegan
A1 Sophia Kitsiou-Tzeli
A1 Maria Tzetis
A1 Gholson J. Lyon
A1 Jesse Gillis
YR 2017
UL http://biorxiv.org/content/early/2017/04/18/128439.abstract
AB In disease expression analysis, looking for shared functional signals from a set of genes which exhibit differential expression is commonplace. We examine the complement as a possibility, that disease genes display “outlier” or unexpected expression relative to broader patterns of functional expression variation. Using six families from the rare TAF1 syndrome disease cohort, we performed family-specific differential expression analyses and find that functional characterization of top candidates enriches for common pathways unlikely to be specifically linked to disease. However, by filtering away common expression changes using known co-expression, we lose all functional enrichment and are left with a small number of outliers characteristic of each proband. Two of these outlier genes are highly recurrent across pedigrees (FDR <2.63e-05) and are the primary commonality among the cohort as a whole. This suggests that systems analysis may be relevant to rare diseases principally as a means of filtering out biological signals unrelated to disease.ROCReceiver Operating CharacteristicAUROCArea Under the ROCDEDifferential ExpressionFDRFalse Discovery Rate (Benjamini-Hochberg correction)FWERFamily-Wise Error Rate (Bonferroni correction)GOGene ontologyCPMCounts Per Million (relative RNA expression based on read counts)