PreTP-2L: identification of therapeutic peptides and their types using two-layer ensemble learning framework

Yan, Ke; Guo, Yichen; Liu, Bin

doi:10.1093/bioinformatics/btad125

Cited by 8 publications

(3 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If the length of the sequence is less than , we pad it with zeros at the end of the sequence, while if the length of the sequence exceeds two sub-sequences with length of 25 from its N-terminal and C-terminal are extracted and concatenated [ 37 ]. We have also tested another sequence truncating strategy, which only extracts the sub-sequence from the sequence beginning side (N-terminal) as [ 5 ] or most of the natural language processing (NLP) tasks [ 38 ] generally do. The performance results listed in Additional file 3 : Table S4 show that the above two truncating strategies are comparable to each other.…”

Section: Methodsmentioning

confidence: 99%

“…Therapeutic peptides play an essential role in human physiology, treatment paradigms, and bio-pharmacy [ 1 – 3 ]. Over the last few decades, peptide-based therapeutics have received a great deal of attention from researchers due to their advantages in drug discovery and design [ 4 , 5 ]. During the epidemic of COVID-19, therapeutic peptides have shown their potential as the agents against SARS-CoV-2 [ 6 – 8 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

TPpred-LE: therapeutic peptide function prediction based on label embedding

Lv,

Yan,

Liu

2023

BMC Biol

Self Cite

View full text Add to dashboard Cite

Background Therapeutic peptides play an essential role in human physiology, treatment paradigms and bio-pharmacy. Several computational methods have been developed to identify the functions of therapeutic peptides based on binary classification and multi-label classification. However, these methods fail to explicitly exploit the relationship information among different functions, preventing the further improvement of the prediction performance. Besides, with the development of peptide detection technology, peptide functions will be more comprehensively discovered. Therefore, it is necessary to explore computational methods for detecting therapeutic peptide functions with limited labeled data. Results In this study, a novel method called TPpred-LE based on Transformer framework was proposed for predicting therapeutic peptide multiple functions, which can explicitly extract the function correlation information by using label embedding methodology and exploit the specificity information based on function-specific classifiers. Besides, we incorporated the multi-label classifier retraining approach (MCRT) into TPpred-LE to detect the new therapeutic functions with limited labeled data. Experimental results demonstrate that TPpred-LE outperforms the other state-of-the-art methods, and TPpred-LE with MCRT is robust for the limited labeled data. Conclusions In summary, TPpred-LE is a function-specific classifier for accurate therapeutic peptide function prediction, demonstrating the importance of the relationship information for therapeutic peptide function prediction. MCRT is a simple but effective strategy to detect functions with limited labeled data.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

TPpred-LE: therapeutic peptide function prediction based on label embedding

Lv,

Yan,

Liu

2023

BMC Biol

Self Cite

View full text Add to dashboard Cite

show abstract

“…Deep learning learns the underlying patterns in large datasets by understanding the laws and layers of sample data to replicate human analytical and learning abilities through artificial intelligence. A range of deep learning-based computational approaches has been used widely in bioinformatics, allowing machines to identify features of protein or peptide sequences and transform these to represent relevant information about the object of interest [ 30–37 ]. Effective sequence representation approaches include soft-alignment mechanism (SSA) [ 38 ], unified representation (UniRep) [ 38–40 ], BiLSTM [ 41 ] and bidirectional encoder representation from transformers (BERT) [ 42 ].…”

Section: Introductionmentioning

confidence: 99%

FEOpti-ACVP: identification of novel anti-coronavirus peptide sequences based on feature engineering and optimization

Jiang,

Pei,

et al. 2024

Briefings in Bioinformatics

View full text Add to dashboard Cite

Anti-coronavirus peptides (ACVPs) represent a relatively novel approach of inhibiting the adsorption and fusion of the virus with human cells. Several peptide-based inhibitors showed promise as potential therapeutic drug candidates. However, identifying such peptides in laboratory experiments is both costly and time consuming. Therefore, there is growing interest in using computational methods to predict ACVPs. Here, we describe a model for the prediction of ACVPs that is based on the combination of feature engineering (FE) optimization and deep representation learning. FEOpti-ACVP was pre-trained using two feature extraction frameworks. At the next step, several machine learning approaches were tested in to construct the final algorithm. The final version of FEOpti-ACVP outperformed existing methods used for ACVPs prediction and it has the potential to become a valuable tool in ACVP drug design. A user-friendly webserver of FEOpti-ACVP can be accessed at http://servers.aibiochem.net/soft/FEOpti-ACVP/.

show abstract