Summary　Alternative　polyadenylation　(APA)　is　now　emerging　as　a　widespread　mechanism　modulated　tissue-specifically,　which　highlights　the　need　to　define　tissue-specific　poly(A)　sites　for　profiling　APA　dynamics　across　tissues.　We　have　developed　an　R　package　called　TSAPA　based　on　the　machine　learning　model　for　identifying　tissue-specific　poly(A)　sites　in　plants.　A　feature　space　including　more　than　200　features　was　assembled　to　specifically　characterize　poly(A)　sites　in　plants.　The　classification　model　in　TSAPA　can　be　customized　by　selecting　desirable　features　or　classifiers.　TSAPA　is　also　capable　of　predicting　tissue-specific　poly(A)　sites　in　unannotated　intergenic　regions.　TSAPA　will　be　a　valuable　addition　to　the　community　for　studying　dynamics　of　APA　in　plants.　Availability　and　implementation　https://github.com/BMILAB/TSAPA　Supplementary　informationSupplementary　dataare　available　at　Bioinformatics　online.　©　The　Author(s)　2018.