GASPIDs Versus Non-GASPIDs - Differentiation Based on Machine Learning Approach

Abstract

Background: Peptidases are a group of enzymes which catalyzes the cleavage of peptide bonds. Around 2-3% of the whole genome codes for proteases and about one third of all known proteases are serine proteases which are divided into 13 clans and 40 families. They are involved in diverse physiological roles such as digestion, coagulation of blood, fibrinolysis, processing of proteins and prohormones, signaling pathways, complement fixation and have a vital role in immune defense system. Based on their functions, they can broadly be divided into two classes; GASPIDs (Granule Associated Serine Peptidases involved in Immune Defense System) and Non-GASPIDs. GASPIDs, in particular are involved in immune associated functions i.e. initiating apoptosis to kill virally infected and cancerous cells, cytokine modulation for generation of inflammatory responses and direct killing of pathogens through phagosomes. Methods: In this study, sequence-based characterization of these two types of serine proteases is performed. We first identified sequences by analyzing multiple online databases as well as by analyzing whole genomes of different species from different orthologous and nonorthologous species. Sequences were identified by devising a distinct criterion to differentiate GASPIDs from Non-GASPIDs. The translated version of these sequences were then subjected to feature extraction. Using these distinctive features, we differentiated GASPIDs from NonGASPIDs by applying multiple supervised machine learning models. Results and Conclusion: Our results show that, among the three classifiers used in this study, SVM classifier coupled with tripeptide as feature method, has shown the best accuracy in classification of sequences as GASPIDs and Non-GASPIDs.

An Efficient Lightweight Hybrid Model with Attention Mechanism for Enhancer Sequence Recognition

“…As much as machine learning-based methods like those mentioned above are capable of delivering good results, it has been shown that deep learning models produce better results without requiring a manual feature extraction operation. Furthermore, machine learning techniques for genomics analysis still require input features that are hand-designed by an individual and extracted from predetermined sequences of input data in order to be able to make decisions [17,18]. It should be noted, however, that convolutional neural networks (CNNs) are able to quickly extract substantial features from a variety of stages.…”

“…In the case that the proposed sequence did not meet all of these criteria, the sequence was referred to as a non-enhancer sequence. The analysis in this study was based on Chou's five-step rule, which has been extensively used in recent studies [17][18][19][20]. As required, the following procedures were followed: (i) for the preparation and testing of the indicator, the large benchmark dataset must be assembled and analyzed; (ii) enhance the significance of genomic sequences by emphasizing a meaningful pattern during the extraction or determination; (iii) develop a classifier capable of identifying these sequences in an effective and accurate manner; (iv) perform various cross-validation techniques on the data; and (v) assemble a web server.…”

Uncovering the evolutionary tail of GZMM: An NSP4 related protease

“…The serine proteases involved in immune defense i.e. apoptosis, cytokine & inflammatory responses and direct killing of pathogens via phagosomes [2,7,14], are more commonly known as GASPIDs (Granules Associated Serine Peptidases of Immune Defense) [2] Whereas, the non-immunological serine proteases involved in digestion, blood coagulation, fibrinolysis, development, and fertilization are recently been referred to as non-GASPIDS [14].…”