Skip to main content
Fig. 3 | BMC Bioinformatics

Fig. 3

From: An unsupervised deep learning framework for predicting human essential genes from population and functional genomic data

Fig. 3

Impact of genomic features on the inference of LOF intolerance. a Association of genomic features with LOF intolerance. We define the contribution score of a genomic feature as the negative weight of the feature in the linear DeepLOF model. The absolute value of a contribution score indicates the strength of association between a feature and LOF intolerance, whereas the sign of a contribution score indicates the direction of association. b DeepLOF automatically adjusts the relative importance of genomic features in a gene length-dependent manner. The x axis represents the expected number of LOF variants. The y axis represents the absolute difference in DeepLOF score between the linear DeepLOF model with genomic features and the counterpart model without genomic features. A higher absolute difference in DeepLOF score indicates a stronger impact of genomic features on the inference of LOF intolerance. Each dot represents a gene. The blue and grey curves represent the fit of the generalized additive model with integrated smoothness and its 95% confidence interval [66]

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us