Category:ECO:0005535 ! machine learning prediction of motif instance evidence

name: machine learning prediction of motif instance evidence
def: "A type of match to sequence model evidence which captures the application of a broad suite of supervised and unsupervised machine learning algorithms trained on known motif instances and applied to the prediction of novel putative instances of a motif in a set of biological sequences." [ECO:SW, PMID:17117497, PMID:2014171]
is_a: ECO:0000202 ! match to sequence model evidence
comment: Machine learning algorithms include, but are not limited to artificial neural networks, random-forest, k-means clustering, hidden Markov models. This suite of methods can be used to identify or verify transcription factor-binding sites in genomic sequences (PMID: 15703297).