Abstract:Regarding to critical roles of oogenesis in formation of ova or unfertilized eggs from the oogonia by mitotic division and subsequent differentiation, the identification of oogenesis-related proteins is of great interest. However, the experimental determination of proteins involved in oogenesis is expensive, time consuming and labor-intensive. Therefore, a new powerful discriminating model is indispensable for classifying oogenesis/non-oogenesis-related proteins with high accuracy and precision. Hereby, for the first time we developed a support vector machine based oogenesis protein prediction method which differentiates oogenesis from non-oogenesis proteins. By means of informative protein physicochemical properties and in addition parameter optimization scheme, our method yields a robust and consistent performance. Our model achieved 87.68% and 84.82% prediction accuracy by five-fold cross validation test for datasets with 90% and 50% identity, respectively. The prediction model was also assessed using the independent dataset and yielded 91.62% and 85.38% prediction accuracy for datasets with 90% and 50% identity, respectively, which further demonstrates the effectiveness of our method. Moreover, by applying 10 different feature weighting methods, the more important protein features for oogenesis/non-oogenesis-related proteins discrimination, including serine and glycine frequency, quasi-sequence-order, pseudo-amino acid composition, distribution and conjoint triad, were determined. The success rates revealed that our model can be considered as a new encouraging and strong model for predicting proteins involved in oogenesis with appropriate performance. To enhance the value of the practical applications of the proposed method, we developed a standalone software for predicting oogenesis candidate proteins called OOgenesis_Pred. This software is the first predictor ever established for identifying oogenesis proteins. We also showed the capability of OOgenesis_Pred by making oogenesis-related proteins prediction for some of the oogenesis candidate proteins. It is anticipated that OOgenesis_Pred will become a powerful tool for future proteomic studies related to oogenesis.

A Computational Model to Identify Fertility-Related Proteins Using Sequence Information

Fertility-LightGBM: A fertility-related protein prediction model by multi-information fusion and light gradient boosting machine

Prediction of Functional Class of Proteins and Peptides Irrespective of Sequence Homology by Support Vector Machines.

Protein deep profile and model predictions for identifying the causal genes of male infertility based on deep learning

Homozygous Loss-of-function Mutations in FSIP2 Cause Male Infertility with Asthenoteratospermia

Application of Eight Machine Learning Algorithms in the Establishment of Infertility and Pregnancy Diagnostic Models: A Comprehensive Analysis of Amino Acid and Carnitine Metabolism

A protein fitness predictive framework based on feature combination and intelligent searching

Multifactor Prediction of Embryo Transfer Outcomes Based on a Machine Learning Algorithm

Identifying Functions of Proteins in Mice With Functional Embedding Features

ApoPred: Identification of Apolipoproteins and Their Subfamilies With Multifarious Features

Predicting Protein Structural Classes for Low-Similarity Sequences by Evaluating Different Features

Advancing the prediction accuracy of protein-protein interactions by utilizing evolutionary information from position-specific scoring matrix and ensemble classifier

Computational prediction of allergenic proteins based on multi-feature fusion

Prediction Of Cyclin Protein Using Two-Step Feature Selection Technique

FertilityOnline: A Straightforward Pipeline for Functional Gene Annotation and Disease Mutation Discovery

Machine-intelligence for developing a potent signature to predict ovarian response to tailor assisted reproduction technology

OOgenesis_Pred: A sequence-based method for predicting oogenesis proteins by six different modes of Chou's pseudo amino acid composition

Simultaneously Identify Three Different Attributes of Proteins by Fusing Their Three Different Modes of Chou'S Pseudo Amino Acid Compositions

Development and validation of a clinical prediction model of fertilization failure during routine IVF cycles

Prediction of Protein Structural Classes Based on Feature Selection Technique.

A model for identification of potential phase-separated proteins based on protein sequence, structure and cellular distribution