afpCOOL: A tool for antifreeze protein prediction

Abstract

Various cold-adapted organisms produce antifreeze proteins (AFPs), which prevent the freezing of cell fluids by inhibiting the growth of ice crystals. AFPs are currently being recognized in various organisms, living in extremely low temperatures. AFPs have several important applications in increasing freeze tolerance of plants, maintaining the tissue in frozen conditions and producing cold-hardy plants by applying transgenic technology. Substantial differences in the sequence and structure of the AFPs, pose a challenge for researchers to identify these proteins. In this paper, we proposed a novel method to identify AFPs, using supportive vector machine (SVM) by incorporating 4 types of features. Results of the two used benchmark datasets, revealed the strength of the proposed method in AFP prediction. According to the results of an independent test setup, our method outperformed the current state-of-the-art methods. In addition, the comparison results of the discrimination power of different feature types revealed that physicochemical descriptors are the most contributing features in AFP detection. This method has been implemented as a stand-alone tool, named afpCOOL, for various operating systems to predict AFPs with a user friendly graphical interface.

Identifying Antifreeze Proteins Based on Key Evolutionary Information

“…However, it is important to note that the PSSMs of our antifreeze proteins were based on comparing sequence similarities with related proteins but not necessarily proteins with antifreeze function. Antifreeze proteins are rare and dissimilar in their sequences, and PSI-BLAST and BLAST have difficulty using an antifreeze protein as the query sequence to search for new antifreeze proteins based on similarity (Kandaswamy et al, 2011;Eslami et al, 2018;Nath and Subbiah, 2018). Thus, some of the sequences that were used to calculate the PSSMs of our antifreeze proteins may have been non-antifreeze protein sequences.…”

“…With the advancements of genome sequencing, a large number of sequenced proteins have been accumulated and need to be functionally annotated. Many auto-annotation tools exist to identify antifreeze proteins, such as TargetFreeze (He et al, 2015), AFP_PSSM (Zhao et al, 2012), CryoProtect (Pratiwi et al, 2017), and afpCOOL (Eslami et al, 2018). However, these tools use too many features ( Table 2), which may often be redundant and lead to overfitting.…”

“…b Results were obtained from a study by Zhao et al (2012). c Results were obtained from a study by Eslami et al (2018). AUC, area under the receiver operating characteristic curve; ACC, accuracy; SN, sensitivity; SP, specificity; MCC, Matthews correlation coefficient.…”

“…However, in spite of similar functions among antifreeze proteins, traditional tools that search for homologous proteins based on sequence similarity, such as Basic Local Alignment Search Tool (BLAST) and Position-Specific Iterative (PSI)-BLAST, perform poorly when attempting to identify antifreeze proteins (Kandaswamy et al, 2011;Eslami et al, 2018;Nath and Subbiah, 2018), because antifreeze proteins exhibit a great diversity among species in their structures and sequence properties. For example, the ice-binding sites in fishes are moderately hydrophobic (Jia and Davies, 2002), while in plants they are mostly hydrophilic (Ramya, 2017).…”

“…He et al (2015) further compared the performances of evolutionary features with two amino acid composition metrics (i.e., amino acid composition and pseudo amino acid composition), and showed that features derived from PSSMs achieved higher performance. Similarly, Yang et al (2015) reported that among various features pertinent to identifying antifreeze proteins, features derived from PSSMs accounted for the largest proportion, though another study showed that physicochemical properties were more important (Eslami et al, 2018). Nevertheless, these results suggest that identifying the evolutionary information underlying the differentiation between antifreeze and non-antifreeze proteins is important for increasing our understanding of proteinice interactions.…”