Improved prediction of malaria degradomes by supervised learning with SVM and profile kernel

Understanding the complex formation of falstatin; an endogenous macromolecular inhibitor of falcipains

Proteolytic activity constitutes a fundamental process essential for the survival of the malaria parasite and is thus highly regulated. Falstatin, a protease inhibitor of Plasmodium falciparum, tightly regulates the activity of cysteine hemoglobinases, falcipain-2 and 3 (FP2, FP3), by inhibiting FP2 through a single surface exposed loop. However, the multimeric nature of falstatin and its interaction with FP2 remained unexplored. Here we report that the N-terminal falstatin region is highly disordered, and needs chaperone activity (heat-shock protein 70, HSP70) for its folding. Protein-protein interaction assays showed a significant interaction between falstatin and HSP70. Further, characterization of the falstatin multimer through a series of biophysical techniques identified the formation of a falstatin decamer, which was extremely thermostable. Computational analysis of the falstatin decamer showed the presence of five falstatin dimers, with each dimer aligned in a head-to-tail orientation. Further, the falstatin C-terminal region was revealed to be primarily involved in the oligomerization process. Stoichiometric analysis of the FP2-falstatin multimer showed the formation of a heterooligomeric complex in a 1:1 ratio, with the participation of ten subunits of each protein. Taken together, our results report a novel protease-inhibitor complex and strengthens our understanding of the regulatory mechanisms of major plasmodium hemoglobinases.

Artificial Intelligence and Machine Learning for Lead-to-Candidate Decision-Making and Beyond

The use of artificial intelligence (AI) and machine learning (ML) in pharmaceutical research and development has to date focused on research: target identification; docking-, fragment-, and motif-based generation of compound libraries; modeling of synthesis feasibility; rank-ordering likely hits according to structural and chemometric similarity to compounds having known activity and affinity to the target(s); optimizing a smaller library for synthesis and high-throughput screening; and combining evidence from screening to support hit-to-lead decisions. Applying AI/ML methods to lead optimization and lead-to-candidate (L2C) decision-making has shown slower progress, especially regarding predicting absorption, distribution, metabolism, excretion, and toxicology properties. The present review surveys reasons why this is so, reports progress that has occurred in recent years, and summarizes some of the issues that remain. Effective AI/ML tools to derisk L2C and later phases of development are important to accelerate the pharmaceutical development process, ameliorate escalating development costs, and achieve greater success rates.

Proteins of the Ciliated Protozoan Parasite Ichthyophthirius multifiliis Identified in Common Carp Skin Mucus

The skin mucus is the fish primary defense barrier protecting from infections via the skin epidermis. In a previous study, we have investigated the proteome of common carp (Cyprinus carpio) skin mucus at two different time points (1 and 9 days) post-exposure to Ichthyophthirius multifiliis. Applying a nano-LC ESI MS/MS technique, we have earlier revealed that the abundance of 44 skin mucus proteins has been differentially regulated including proteins associated with host immune responses and wound healing. Herein, in skin mucus samples, we identified six proteins of I. multifiliis associated with the skin mucus in common carp. Alpha and beta tubulins were detected in addition to the elongation factor alpha, 26S proteasome regulatory subunit, 26S protease regulatory subunit 6B, and heat shock protein 90. The identified proteins are likely involved in motility, virulence, and general stress during parasite growth and development after parasite attachment and invasion. Two KEGG pathways, phagosome and proteasome, were identified among these parasite proteins, mirroring the proteolytic and phagocytic activities of this parasite during host invasion, growth, and development, which represent a plausible host invasion strategy of this parasite. The results obtained from this study can support revealing molecular aspects of the interplay between carp and I. multifiliis and may help us understand the I. multifiliis invasion strategy at the skin mucus barrier. The data may advance the development of novel drugs, vaccines, and diagnostics suitable for the management and prevention of ichthyophthiriosis in fish.

Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Proteomic Analysis of Plasmodium Merosomes: The Link between Liver and Blood Stages in Malaria

The pre-erythrocytic liver stage of the malaria parasite, comprising sporozoites and the liver stages into which they develop, remains one of the least understood parts of the lifecycle, in part owing to the low numbers of parasites. Nonetheless, it is recognized as an important target for antimalarial drugs and vaccines. Here we provide the first proteomic analysis of merosomes, which define the final phase of the liver stage and are responsible for initiating the blood stage of infection. We identify a total of 1879 parasite proteins, and a core set of 1188 proteins quantitatively detected in every biological replicate, providing an extensive picture of the protein repertoire of this stage. This unique data set will allow us to explore key questions about the biology of merosomes and hepatic merozoites.

Proteomic analysis of differentially expressed proteins in the two developmental stages of Ichthyophthirius multifiliis

Ichthyophthirius is a severe disease of farmed freshwater fish caused by the parasitic ciliate Ichthyophthirius multifiliis (Ich). This disease can lead to considerable economic loss, but the protein profiles in different developmental stages of the parasite remain unknown. In the present study, proteins from trophonts and theronts of Ich were identified by isobaric tags for relative and absolute quantitation (iTRAQ). A total of 2300 proteins were identified in the two developmental stages, of which 1520 proteins were differentially expressed. Among them, 84 proteins were uniquely expressed in the theronts stage, while 656 proteins were expressed only in trophonts. The differentially expressed proteins were catalogued (assorted) to various functions of Ich life cycle, including biological process, cellular component, and molecular function that occur at distinct stages. Using a 1.5-fold change in expression as a physiologically significant benchmark, a lot of differentially expressed proteins were reliably quantified by iTRAQ analysis. Two hundred forty upregulated and 57 downregulated proteins in the trophonts stage were identified as compared with theronts. The identified proteins were involved in various functions of the I. multifiliis life cycle, including binding, catalytic activity, structural molecule activity, and transporter activity. Further investigation of the transcriptional levels of periplasmic immunogenic protein, transketolase, zinc finger, isocitrate dehydrogenase, etc., from the different protein profiles using quantitative RT-PCR showed identical results to the iTRAQ analysis. This work provides an effective resource to further our understanding of Ich biology, and lays the groundwork for the identification of potential drug targets and vaccines candidates for the control of this devastating fish pathogen.

Blocking Plasmodium falciparum development via dual inhibition of hemoglobin degradation and the ubiquitin proteasome system by MG132

Among key potential drug target proteolytic systems in the malaria parasite Plasmodium falciparum are falcipains, a family of hemoglobin-degrading cysteine proteases, and the ubiquitin proteasomal system (UPS), which has fundamental importance in cellular protein turnover. Inhibition of falcipains blocks parasite development, primarily due to inhibition of hemoglobin degradation that serves as a source of amino acids for parasite growth. Falcipains prefer P2 leucine in substrates and peptides, and their peptidyl inhibitors with leucine at the P2 position show potent antimalarial activity. The peptidyl inhibitor MG132 (Z-Leu-Leu-Leu-CHO) is a widely used proteasome inhibitor, which also has P2 leucine, and has also been shown to inhibit parasite development. However, the antimalarial targets of MG132 are unclear. We investigated whether MG132 blocks malaria parasite development by inhibiting hemoglobin degradation and/or by targeting the UPS. P. falciparum was cultured with inhibitors of the UPS (MG132, epoxomicin, and lactacystin) or falcipains (E64), and parasites were assessed for morphologies, extent of hemoglobin degradation, and accumulation of ubiquitinated proteins. MG132, like E64 and unlike epoxomicin or lactacystin, blocked parasite development, with enlargement of the food vacuole and accumulation of undegraded hemoglobin, indicating inhibition of hemoglobin degradation by MG132, most likely due to inhibition of hemoglobin-degrading falcipain cysteine proteases. Parasites cultured with epoxomicin or MG132 accumulated ubiquitinated proteins to a significantly greater extent than untreated or E64-treated parasites, indicating that MG132 inhibits the parasite UPS as well. Consistent with these findings, MG132 inhibited both cysteine protease and UPS activities present in soluble parasite extracts, and it strongly inhibited recombinant falcipains. MG132 was highly selective for inhibition of P. falciparum (IC₅₀ 0.0476 µM) compared to human peripheral blood mononuclear cells (IC₅₀ 10.8 µM). Thus, MG132 inhibits two distinct proteolytic systems in P. falciparum, and it may serve as a lead molecule for development of dual-target inhibitors of malaria parasites.

Module-based subnetwork alignments reveal novel transcriptional regulators in malaria parasite Plasmodium falciparum

Background

Malaria causes over one million deaths annually, posing an enormous health and economic burden in endemic regions. The completion of genome sequencing of the causative agents, a group of parasites in the genus Plasmodium, revealed potential drug and vaccine candidates. However, genomics-driven target discovery has been significantly hampered by our limited knowledge of the cellular networks associated with parasite development and pathogenesis. In this paper, we propose an approach based on aligning neighborhood PPI subnetworks across species to identify network components in the malaria parasite P. falciparum.

Results

Instead of only relying on sequence similarities to detect functional orthologs, our approach measures the conservation between the neighborhood subnetworks in protein-protein interaction (PPI) networks in two species, P. falciparum and E. coli. 1,082 P. falciparum proteins were predicted as functional orthologs of known transcriptional regulators in the E. coli network, including general transcriptional regulators, parasite-specific transcriptional regulators in the ApiAP2 protein family, and other potential regulatory proteins. They are implicated in a variety of cellular processes involving chromatin remodeling, genome integrity, secretion, invasion, protein processing, and metabolism.

Conclusions

In this proof-of-concept study, we demonstrate that a subnetwork alignment approach can reveal previously uncharacterized members of the subnetworks, which opens new opportunities to identify potential therapeutic targets and provide new insights into parasite biology, pathogenesis and virulence. This approach can be extended to other systems, especially those with poor genome annotation and a paucity of knowledge about cellular networks.

Comparative Genomics and Systems Biology of Malaria Parasites Plasmodium

Malaria is a serious infectious disease that causes over one million deaths yearly. It is caused by a group of protozoan parasites in the genus Plasmodium. No effective vaccine is currently available and the elevated levels of resistance to drugs in use underscore the pressing need for novel antimalarial targets. In this review, we survey omics centered developments in Plasmodium biology, which have set the stage for a quantum leap in our understanding of the fundamental processes of the parasite life cycle and mechanisms of drug resistance and immune evasion.

Proteases in malaria parasites - a phylogenomic perspective

Malaria continues to be one of the most devastating global health problems due to the high morbidity and mortality it causes in endemic regions. The search for new antimalarial targets is of high priority because of the increasing prevalence of drug resistance in malaria parasites. Malarial proteases constitute a class of promising therapeutic targets as they play important roles in the parasite life cycle and it is possible to design and screen for specific protease inhibitors. In this mini-review, we provide a phylogenomic overview of malarial proteases. An evolutionary perspective on the origin and divergence of these proteases will provide insights into the adaptive mechanisms of parasite growth, development, infection, and pathogenesis.B

Protease-associated cellular networks in malaria parasite Plasmodium falciparum

Background

Malaria continues to be one of the most severe global infectious diseases, responsible for 1-2 million deaths yearly. The rapid evolution and spread of drug resistance in parasites has led to an urgent need for the development of novel antimalarial targets. Proteases are a group of enzymes that play essential roles in parasite growth and invasion. The possibility of designing specific inhibitors for proteases makes them promising drug targets. Previously, combining a comparative genomics approach and a machine learning approach, we identified the complement of proteases (degradome) in the malaria parasite Plasmodium falciparum and its sibling species [1-3], providing a catalog of targets for functional characterization and rational inhibitor design. Network analysis represents another route to revealing the role of proteins in the biology of parasites and we use this approach here to expand our understanding of the systems involving the proteases of P. falciparum.

Results

We investigated the roles of proteases in the parasite life cycle by constructing a network using protein-protein association data from the STRING database [4], and analyzing these data, in conjunction with the data from protein-protein interaction assays using the yeast 2-hybrid (Y2H) system [5], blood stage microarray experiments [6-8], proteomics [9-12], literature text mining, and sequence homology analysis. Seventy-seven (77) out of 124 predicted proteases were associated with at least one other protein, constituting 2,431 protein-protein interactions (PPIs). These proteases appear to play diverse roles in metabolism, cell cycle regulation, invasion and infection. Their degrees of connectivity (i.e., connections to other proteins), range from one to 143. The largest protease-associated sub-network is the ubiquitin-proteasome system which is crucial for protein recycling and stress response. Proteases are also implicated in heat shock response, signal peptide processing, cell cycle progression, transcriptional regulation, and signal transduction networks.

Conclusions

Our network analysis of proteases from P. falciparum uses a so-called guilt-by-association approach to extract sets of proteins from the proteome that are candidates for further study. Novel protease targets and previously unrecognized members of the protease-associated sub-systems provide new insights into the mechanisms underlying parasitism, pathogenesis and virulence.

Comparative genomics of the pathogenic ciliate Ichthyophthirius multifiliis, its free-living relatives and a host species provide insights into adoption of a parasitic lifestyle and prospects for disease control

Background

Ichthyophthirius multifiliis, commonly known as Ich, is a highly pathogenic ciliate responsible for 'white spot', a disease causing significant economic losses to the global aquaculture industry. Options for disease control are extremely limited, and Ich's obligate parasitic lifestyle makes experimental studies challenging. Unlike most well-studied protozoan parasites, Ich belongs to a phylum composed primarily of free-living members. Indeed, it is closely related to the model organism Tetrahymena thermophila. Genomic studies represent a promising strategy to reduce the impact of this disease and to understand the evolutionary transition to parasitism.

Results

We report the sequencing, assembly and annotation of the Ich macronuclear genome. Compared with its free-living relative T. thermophila, the Ich genome is reduced approximately two-fold in length and gene density and three-fold in gene content. We analyzed in detail several gene classes with diverse functions in behavior, cellular function and host immunogenicity, including protein kinases, membrane transporters, proteases, surface antigens and cytoskeletal components and regulators. We also mapped by orthology Ich's metabolic pathways in comparison with other ciliates and a potential host organism, the zebrafish Danio rerio.

Conclusions

Knowledge of the complete protein-coding and metabolic potential of Ich opens avenues for rational testing of therapeutic drugs that target functions essential to this parasite but not to its fish hosts. Also, a catalog of surface protein-encoding genes will facilitate development of more effective vaccines. The potential to use T. thermophila as a surrogate model offers promise toward controlling 'white spot' disease and understanding the adaptation to a parasitic lifestyle.

Gene ontology based transfer learning for protein subcellular localization

BACKGROUND

Prediction of protein subcellular localization generally involves many complex factors, and using only one or two aspects of data information may not tell the true story. For this reason, some recent predictive models are deliberately designed to integrate multiple heterogeneous data sources for exploiting multi-aspect protein feature information. Gene ontology, hereinafter referred to as GO, uses a controlled vocabulary to depict biological molecules or gene products in terms of biological process, molecular function and cellular component. With the rapid expansion of annotated protein sequences, gene ontology has become a general protein feature that can be used to construct predictive models in computational biology. Existing models generally either concatenated the GO terms into a flat binary vector or applied majority-vote based ensemble learning for protein subcellular localization, both of which can not estimate the individual discriminative abilities of the three aspects of gene ontology.

RESULTS

In this paper, we propose a Gene Ontology Based Transfer Learning Model (GO-TLM) for large-scale protein subcellular localization. The model transfers the signature-based homologous GO terms to the target proteins, and further constructs a reliable learning system to reduce the adverse affect of the potential false GO terms that are resulted from evolutionary divergence. We derive three GO kernels from the three aspects of gene ontology to measure the GO similarity of two proteins, and derive two other spectrum kernels to measure the similarity of two protein sequences. We use simple non-parametric cross validation to explicitly weigh the discriminative abilities of the five kernels, such that the time & space computational complexities are greatly reduced when compared to the complicated semi-definite programming and semi-indefinite linear programming. The five kernels are then linearly merged into one single kernel for protein subcellular localization. We evaluate GO-TLM performance against three baseline models: MultiLoc, MultiLoc-GO and Euk-mPLoc on the benchmark datasets the baseline models adopted. 5-fold cross validation experiments show that GO-TLM achieves substantial accuracy improvement against the baseline models: 80.38% against model Euk-mPLoc 67.40% with 12.98% substantial increase; 96.65% and 96.27% against model MultiLoc-GO 89.60% and 89.60%, with 7.05% and 6.67% accuracy increase on dataset MultiLoc plant and dataset MultiLoc animal, respectively; 97.14%, 95.90% and 96.85% against model MultiLoc-GO 83.70%, 90.10% and 85.70%, with accuracy increase 13.44%, 5.8% and 11.15% on dataset BaCelLoc plant, dataset BaCelLoc fungi and dataset BaCelLoc animal respectively. For BaCelLoc independent sets, GO-TLM achieves 81.25%, 80.45% and 79.46% on dataset BaCelLoc plant holdout, dataset BaCelLoc plant holdout and dataset BaCelLoc animal holdout, respectively, as compared against baseline model MultiLoc-GO 76%, 60.00% and 73.00%, with accuracy increase 5.25%, 20.45% and 6.46%, respectively.

CONCLUSIONS

Since direct homology-based GO term transfer may be prone to introducing noise and outliers to the target protein, we design an explicitly weighted kernel learning system (called Gene Ontology Based Transfer Learning Model, GO-TLM) to transfer to the target protein the known knowledge about related homologous proteins, which can reduce the risk of outliers and share knowledge between homologous proteins, and thus achieve better predictive performance for protein subcellular localization. Cross validation and independent test experimental results show that the homology-based GO term transfer and explicitly weighing the GO kernels substantially improve the prediction performance.

Core genome components and lineage specific expansions in malaria parasites plasmodium

Background

The increasing resistance of Plasmodium, the malaria parasites, to multiple commonly used drugs has underscored the urgent need to develop effective antimalarial drugs and vaccines. The new direction of genomics-driven target discovery has become possible with the completion of parasite genome sequencing, which can lead us to a better understanding of how the parasites develop the genetic variability that is associated with their response to environmental challenges and other adaptive phenotypes.

Results

We present the results of a comprehensive analysis of the genomes of six Plasmodium species, including two species that infect humans, one that infects monkeys, and three that infect rodents. The core genome shared by all six species is composed of 3,351 genes, which make up about 22%-65% of the genome repertoire. These components play important roles in fundamental functions as well as in parasite-specific activities. We further investigated the distribution and features of genes that have been expanded in specific Plasmodium lineage(s). Abundant duplicate genes are present in the six species, with 5%-9% of the whole genomes composed lineage specific radiations. The majority of these gene families are hypothetical proteins with unknown functions; a few may have predicted roles such as antigenic variation.

Conclusions

The core genome components in the malaria parasites have functions ranging from fundamental biological processes to roles in the complex networks that sustain the parasite-specific lifestyles appropriate to different hosts. They represent the minimum requirement to maintain a successful life cycle that spans vertebrate hosts and mosquito vectors. Lineage specific expansions (LSEs) have given rise to abundant gene families in Plasmodium. Although the functions of most families remain unknown, these LSEs could reveal components in parasite networks that, by their enhanced genetic variability, can contribute to pathogenesis, virulence, responses to environmental challenges, or interesting phenotypes.

Amino acid classification based spectrum kernel fusion for protein subnuclear localization

BACKGROUND

Prediction of protein localization in subnuclear organelles is more challenging than general protein subcelluar localization. There are only three computational models for protein subnuclear localization thus far, to the best of our knowledge. Two models were based on protein primary sequence only. The first model assumed homogeneous amino acid substitution pattern across all protein sequence residue sites and used BLOSUM62 to encode k-mer of protein sequence. Ensemble of SVM based on different k-mers drew the final conclusion, achieving 50% overall accuracy. The simplified assumption did not exploit protein sequence profile and ignored the fact of heterogeneous amino acid substitution patterns across sites. The second model derived the PsePSSM feature representation from protein sequence by simply averaging the profile PSSM and combined the PseAA feature representation to construct a kNN ensemble classifier Nuc-PLoc, achieving 67.4% overall accuracy. The two models based on protein primary sequence only both achieved relatively poor predictive performance. The third model required that GO annotations be available, thus restricting the model's applicability.

METHODS

In this paper, we only use the amino acid information of protein sequence without any other information to design a widely-applicable model for protein subnuclear localization. We use K-spectrum kernel to exploit the contextual information around an amino acid and the conserved motif information. Besides expanding window size, we adopt various amino acid classification approaches to capture diverse aspects of amino acid physiochemical properties. Each amino acid classification generates a series of spectrum kernels based on different window size. Thus, (I) window expansion can capture more contextual information and cover size-varying motifs; (II) various amino acid classifications can exploit multi-aspect biological information from the protein sequence. Finally, we combine all the spectrum kernels by simple addition into one single kernel called SpectrumKernel+ for protein subnuclear localization.

RESULTS

We conduct the performance evaluation experiments on two benchmark datasets: Lei and Nuc-PLoc. Experimental results show that SpectrumKernel+ achieves substantial performance improvement against the previous model Nuc-PLoc, with overall accuracy 83.47% against 67.4%; and 71.23% against 50% of Lei SVM Ensemble, against 66.50% of Lei GO SVM Ensemble.

CONCLUSION

The method SpectrumKernel+ can exploit rich amino acid information of protein sequence by embedding into implicit size-varying motifs the multi-aspect amino acid physiochemical properties captured by amino acid classification approaches. The kernels derived from diverse amino acid classification approaches and different sizes of k-mer are summed together for data integration. Experiments show that the method SpectrumKernel+ significantly outperforms the existing models for protein subnuclear localization.