In metazoans, lysosomes are the center for the degradation of macromolecules and play a key role in a variety of cellular processes, such as autophagy, exocytosis and membrane repair. Defects of lysosomal pathways are associated with lysosomal storage disorders and with several late onset neurodegenerative diseases. We recently discovered the CLEAR (Coordinated Lysosomal Expression and Regulation) gene network and its master gene transcription factor EB (TFEB), which regulates lysosomal biogenesis and function. Here, we used a combination of genomic approaches, including ChIP-seq (sequencing of chromatin immunoprecipitate) analysis, profiling of TFEB-mediated transcriptional induction, genome-wide mapping of TFEB target sites and recursive expression meta-analysis of TFEB targets, to identify 471 TFEB direct targets that represent essential components of the CLEAR network. This analysis revealed a comprehensive system regulating the expression, import and activity of lysosomal enzymes that control the degradation of proteins, glycosaminoglycans, sphingolipids and glycogen. Interestingly, the CLEAR network appears to be involved in the regulation of additional lysosome-associated processes, including autophagy, exo- and endocytosis, phagocytosis and immune response. Furthermore, non-lysosomal enzymes involved in the degradation of essential proteins such as hemoglobin and chitin are also part of the CLEAR network. Finally, we identified nine novel lysosomal proteins by using the CLEAR network as a tool for prioritizing candidates. This study provides potential therapeutic targets to modulate cellular clearance in a variety of disease conditions.

Characterization of the CLEAR network reveals an integrated control of cellular clearance pathways

Proteins localized to various cellular and subcellular membranes play pivotal roles in numerous cellular activities. Accordingly, in eukaryotic cells, the biogenesis of organellar proteins is an essential process requiring their correct localization among various cellular and subcellular membranes. Localization of these proteins is determined by either cotranslational or posttranslational mechanisms, depending on the final destination. However, it is not fully understood how the targeting specificity of membrane proteins is determined in plant cells. Here, we investigate the mechanism by which signal-anchored (SA) proteins are differentially targeted to the endoplasmic reticulum (ER) or endosymbiotic organelles using in vivo targeting, subcellular fractionation, and bioinformatics approaches. For targeting SA proteins to endosymbiotic organelles, the C-terminal positively charged region (CPR) flanking the transmembrane domain (TMD) is necessary but not sufficient. The hydrophobicity of the TMD in CPR-containing proteins also plays a critical role in determining targeting specificity; TMDs with a hydrophobicity value >0.4 on the Wimley and White scale are targeted primarily to the ER, whereas TMDs with lower values are targeted to endosymbiotic organelles. Based on these data, we propose that the CPR and the hydrophobicity of the TMD play a critical role in determining the targeting specificity between the ER and endosymbiotic organelles.

/pdf/both-the-hydrophobicity-and-a-positively-charged-region-fwdbowveh0.pdf

Both the Hydrophobicity and a Positively Charged Region Flanking the C-Terminal Region of the Transmembrane Domain of Signal-Anchored Proteins Play Critical Roles in Determining Their Targeting Specificity to the Endoplasmic Reticulum or Endosymbiotic Organelles in Arabidopsis Cells

Apoptosis proteins are very important for understanding the mechanism of programmed cell death. Obtaining information on subcellular location of apoptosis proteins is very helpful to understand the apoptosis mechanism. In this paper, based on amino acid substitution matrix and auto covariance transformation, we introduce a new sequence-based model, which not only quantitatively describes the differences between amino acids, but also partially incorporates the sequence-order information. This method is applied to predict the apoptosis proteins’ subcellular location of two widely used datasets by the support vector machine classifier. The results obtained by jackknife test are quite promising, indicating that the proposed method might serve as a potential and efficient prediction model for apoptosis protein subcellular location prediction.

Predicting subcellular location of apoptosis proteins with pseudo amino acid composition: approach from amino acid substitution matrix and auto covariance transformation.

Background: Computational prediction of protein subcellular localization can greatly help to elucidate its functions. Despite the existence of dozens of protein localization prediction algorithms, the prediction accuracy and coverage are still low. Several ensemble algorithms have been proposed to improve the prediction performance, which usually include as many as 10 or more individual localization algorithms. However, their performance is still limited by the running complexity and redundancy among individual prediction algorithms. Results: This paper proposed a novel method for rational design of minimalist ensemble algorithms for practical genome-wide protein subcellular localization prediction. The algorithm is based on combining a feature selection based filter and a logistic regression classifier. Using a novel concept of contribution scores, we analyzed issues of algorithm redundancy, consensus mistakes, and algorithm complementarity in designing ensemble algorithms. We applied the proposed minimalist logistic regression (LR) ensemble algorithm to two genome-wide datasets of Yeast and Human and compared its performance with current ensemble algorithms. Experimental results showed that the minimalist ensemble algorithm can achieve high prediction accuracy with only 1/3 to 1/2 of individual predictors of current ensemble algorithms, which greatly reduces computational complexity and running time. It was found that the high performance ensemble algorithms are usually composed of the predictors that together cover most of available features. Compared to the best individual predictor, our ensemble algorithm improved the prediction accuracy from AUC score of 0.558 to 0.707 for the Yeast dataset and from 0.628 to 0.646 for the Human dataset. Compared with popular weighted voting based ensemble algorithms, our classifier-based ensemble algorithms achieved much better performance without suffering from inclusion of too many individual predictors. Conclusions: We proposed a method for rational design of minimalist ensemble algorithms using feature selection and classifiers. The proposed minimalist ensemble algorithm based on logistic regression can achieve equal or better prediction performance while using only half or one-third of individual predictors compared to other ensemble algorithms. The results also suggested that meta-predictors that take advantage of a variety of features by combining individual predictors tend to achieve the best performance. The LR ensemble server and related benchmark datasets are available at http://mleg.cse.sc.edu/LRensemble/cgi-bin/predict.cgi.

/pdf/minimalist-ensemble-algorithms-for-genome-wide-protein-5fqw2okl1p.pdf

Minimalist ensemble algorithms for genome-wide protein localization prediction

Structural genomics (SG) is an international effort that aims at solving three-dimensional shapes of important biological macro-molecules with primary focus on proteins. One of the main bottlenecks in SG is the ability to produce diffraction quality crystals for X-ray crystallography based protein structure determination. SG pipelines allow for certain flexibility in target selection which motivates development of insilico methods for sequence-based prediction/ assessment of the protein crystallization propensity. We overview existing SG databanks that are used to derive these predictive models and we discuss analytical results concerning protein sequence properties that were discovered to correlate with the ability to form crystals. We also contrast and empirically compare modern sequence-based predictors of crystallization propensity including OB-Score, ParCrys, XtalPred and CRYSTALP2. Our analysis shows that these methods provide useful and complimentary predictions. Although their average accuracy is similar at around 70%, we show that application of a simple majority-vote based ensemble improves accuracy to almost 74%. The best improvements are achieved by combining XtalPred with CRYSTALP2 while OB-Score and ParCrys methods overlap to a larger extend, although they still complement the other two predictors. We also demonstrate that 90% of the protein chains can be correctly predicted by at least one of these methods, which suggests that more accurate ensembles could be built in the future. We believe that current protein crystallization propensity predictors could provide useful input for the target selection procedures utilized by the SG centers.

/pdf/sequence-based-protein-crystallization-propensity-prediction-2c4qc58l7i.pdf

Sequence-Based Protein Crystallization Propensity Prediction for Structural Genomics: Review and Comparative Analysis

In the past decade, many automated prediction methods for the subcellular localization of proteins have been proposed, utilizing a wide range of principles and learning approaches. Based on an experimental evaluation of different methods and their theoretical properties, we propose to combine a well-balanced set of existing approaches to new, ensemble-based prediction methods. The experimental evaluation shows that our ensembles improve substantially over the underlying base methods.

/pdf/supervised-ensembles-of-prediction-methods-for-subcellular-32erbqe5zb.pdf

Supervised ensembles of prediction methods for subcellular localization.

When considering the prediction of a structural class for a protein as a classification problem, usually a classifier is based on a feature vector x ∈ R n , where the features represent certain attributes of the primary sequence or derived properties (e.g., the predicted secondary structure) of a given protein. Since the structure of a protein (i.e., its native conformation) is stable only under specific environmental conditions, it is commonly accepted to assume proteins being evolutionarily adapted to specific subcellular localizations and according to their physicochem- ical environment. Our statistical evaluation shows a strong correlation between the subcellular localization of proteins and their structural class. The correlation is strong enough to allow for a classification of proteins into their structural class solely based on information regarding the subcellular localization. We conclude that knowledge regarding the subcellular localization of proteins can be useful as a feature for the structural classification of proteins.

http://www.jucs.org/jucs_16_5/investigating_a_correlation_between/jucs_16_05_0604_0621_assfalg.pdf

Investigating a Correlation between Subcellular Localization and Fold of Proteins

/pdf/supervised-ensembles-of-prediction-methods-for-subcellular-32m99yuypl.pdf

Jing Gong

Papers

Supervised ensembles of prediction methods for subcellular localization.

Investigating a Correlation between Subcellular Localization and Fold of Proteins

Supervised Ensembles of Prediction Methods for Subcellular Localization.