Positive-unlabeled learning in bioinformatics and computational biology: a brief review

Abstract

Conventional supervised binary classification algorithms have been widely applied to address significant research questions using biological and biomedical data. This classification scheme requires two fully labeled classes of data (e.g. positive and negative samples) to train a classification model. However, in many bioinformatics applications, labeling data is laborious, and the negative samples might be potentially mislabeled due to the limited sensitivity of the experimental equipment. The positive unlabeled (PU) learning scheme was therefore proposed to enable the classifier to learn directly from limited positive samples and a large number of unlabeled samples (i.e. a mixture of positive or negative samples). To date, several PU learning algorithms have been developed to address various biological questions, such as sequence identification, functional site characterization and interaction prediction. In this paper, we revisit a collection of 29 state-of-the-art PU learning bioinformatic applications to address various biological questions. Various important aspects are extensively discussed, including PU learning methodology, biological application, classifier design and evaluation strategy. We also comment on the existing issues of PU learning and offer our perspectives for the future development of PU learning applications. We anticipate that our work serves as an instrumental guideline for a better understanding of the PU learning framework in bioinformatics and further developing next-generation PU learning frameworks for critical biological applications.

Computational analysis and prediction of PE_PGRS proteins using machine learning

“…They are usually performed worse, especially for those proteins with low sequence similarity with the proteins in the search library. Machine learning combined with extensive sequence feature engineering techniques has been successfully used in many bioinformatics topics [ [21] , [22] , [23] , [24] , [25] , [26] , [27] , [28] , [29] , [30] , [31] , [32] , [69] ], and provide an alternative efficient and accurate strategy to study these enigmatic proteins. As such, we are highly motivated to leverage cutting-edge machine learning techniques to develop computational approaches to identify the PE_PGRS proteins rapidly and accurately.…”

NNAN: Nearest Neighbor Attention Network to Predict Drug–Microbe Associations

“…To evaluate the performance of these methods, we regard the known DMA pairs as positive samples and unlabeled DMA pairs as negative samples ( Peng et al, 2020 ; Li et al, 2022 ). We set up a 5-fold cross-validation scenario in which we randomly divide positive samples and negative samples into five groups, respectively.…”

PCfun: a hybrid computational framework for systematic characterization of protein complex function

“…Another noteworthy issue is the negative data for training the supervised RF model of PCfun. This problem has been brought to attention and discussed in our previous studies [ 57 , 58 ]. In this study, similar issues occurred when constructing the negative PC–GO associations in the training datasets (‘ Supplementary Methods ’).…”

“…Theoretically, there would be many negative protein complex-GO pairs, some of which might be mislabeled. Compared to the traditional supervised machine-learning models, positive-unlabeled learning [ 57 , 59 ] only requires positive and unlabeled (i.e. either positive or negative) training samples to build reliable predictors with competitive prediction performance and, therefore, can be considered as an alternative option for tackling this issue.…”