Shigemitsu, Shunsuke Cao, Wei Terada, Tohru Shimizu, Kentaro Additional file 3 of Development of a prediction system for tail-anchored proteins Figure S3. Lengths and hydrophobicities of sequences that were not predicted. Upper: Sequence lengths. The gray bars represent all TA proteins and red portions indicating prediction failures. Lower: Hydrophobicity scores of unsuccessfully predicted sequences. Linear smoothing was applied to Kyte–Doolittle hydrophobicity scores. Each row corresponds to a sequence, and each column represents a residue. The sequences are listed in order of hydrophobicity between the C-terminus and the 50th residue (sequences in the upper rows are more hydrophobic). (TIF 1658 kb) TA proteins;HMMs;Membrane proteins;Prediction;Machine learning 2016-09-15
    https://springernature.figshare.com/articles/figure/Additional_file_3_of_Development_of_a_prediction_system_for_tail-anchored_proteins/4364675
10.6084/m9.figshare.c.3611723_D2.v1