Small open reading frames (smORFs) encoding proteins less than 100 amino acids (aa) are known to be important regulators of key cellular processes. However, their computational identification remains a challenge. Based on a comprehensive analysis of known prokaryotic small ORFs, we have developed the ProsmORF-pred resource which uses a machine learning (ML)-based method for prediction of smORFs in the prokaryotic genome sequences. ProsmORF-pred consists of two ML models, one for initiation site recognition in nucleic acid sequences upstream of putative start codons and the other uses translated amino acid sequences to decipher functional protein like sequences. The nucleotide sequence-based initiation site recognition model has been trained using longer ORFs (>100 aa) in the same genome while the ML model for identification of protein like sequences has been trained using annotated smORFs from Escherichia coli. Comprehensive benchmarking of ProsmORF-pred reveals that its performance is comparable to other state-of-the-art approaches on the annotated smORF set derived from 32 prokaryotic genomes. Its performance is distinctly superior to other tools like PRODIGAL and RANSEPS for prediction of newly identified smORFs which have a length range of 10-30 aa, where prediction of smORFs has been a major challenge. Apart from identification of smORFs in genomic sequences, ProsmORF-pred can also aid in functional annotation of the predicted smORFs based on sequence similarity and genomic neighbourhood similarity searches in ProsmORFDB, a well-curated database of known smORFs. ProsmORF-pred along with its backend database ProsmORFDB is available as a user-friendly web server (http://www.nii.ac.in/prosmorfpred.html).

Even though several in silico tools are available for prediction of the phosphorylation sites for mammalian, yeast or plant proteins, currently no software is available for predicting phosphosites for Plasmodium proteins. However, the availability of significant amount of phospho-proteomics data during the last decade and advances in machine learning (ML) algorithms have opened up the opportunities for deciphering phosphorylation patterns of plasmodial system and developing ML-based phosphosite prediction tools for Plasmodium. We have developed Pf-Phospho, an ML-based method for prediction of phosphosites by training Random Forest classifiers using a large data set of 12 096 phosphosites of Plasmodium falciparum and Plasmodium bergei. Of the 12 096 known phosphosites, 75% of sites have been used for training/validation of the classifier, while remaining 25% have been used as completely unseen test data for blind testing. It is encouraging to note that Pf-Phospho can predict the kinase-independent phosphosites with 84% sensitivity, 75% specificity and 78% precision. In addition, it can also predict kinase-specific phosphosites for five plasmodial kinases-PfPKG, Plasmodium falciparum, PfPKA, PfPK7 and PbCDPK4 with high accuracy. Pf-Phospho (http://www.nii.ac.in/pfphospho.html) outperforms other widely used phosphosite prediction tools, which have been trained using mammalian phosphoproteome data. It also has been integrated with other widely used resources such as PlasmoDB, MPMP, Pfam and recently available ML-based predicted structures by AlphaFold2. Currently, Pf-phospho is the only bioinformatics resource available for ML-based prediction of phospho-signaling networks of Plasmodium and is a user-friendly platform for integrative analysis of phospho-signaling along with metabolic and protein-protein interaction networks.

Small molecule modulators of protein-protein interactions (PPIs) are being pursued as novel anticancer, antiviral and antimicrobial drug candidates. We have utilized a large data set of experimentally validated PPI modulators and developed machine learning classifiers for prediction of new small molecule modulators of PPI. Our analysis reveals that using random forest (RF) classifier, general PPI Modulators independent of PPI family can be predicted with ROC-AUC higher than 0.9, when training and test sets are generated by random split. The performance of the classifier on data sets very different from those used in training has also been estimated by using different state of the art protocols for removing various types of bias in division of data into training and test sets. The family-specific PPIM predictors developed in this work for 11 clinically important PPI families also have prediction accuracies of above 90% in majority of the cases. All these ML-based predictors have been implemented in a freely available software named SMMPPI for prediction of small molecule modulators for clinically relevant PPIs like RBD:hACE2, Bromodomain_Histone, BCL2-Like_BAX/BAK, LEDGF_IN, LFA_ICAM, MDM2-Like_P53, RAS_SOS1, XIAP_Smac, WDR5_MLL1, KEAP1_NRF2 and CD4_gp120. We have identified novel chemical scaffolds as inhibitors for RBD_hACE PPI involved in host cell entry of SARS-CoV-2. Docking studies for some of the compounds reveal that they can inhibit RBD_hACE2 interaction by high affinity binding to interaction hotspots on RBD. Some of these new scaffolds have also been found in SARS-CoV-2 viral growth inhibitors reported recently; however, it is not known if these molecules inhibit the entry phase.

MOTIVATION: Even though genome mining tools have successfully identified large numbers of non-ribosomal peptide synthetase (NRPS) and polyketide synthase (PKS) biosynthetic gene clusters (BGCs) in bacterial genomes, currently no tool can predict the chemical structure of the secondary metabolites biosynthesized by these BGCs. Lack of algorithms for predicting complex macrocyclization patterns of linear PK/NRP biosynthetic intermediates has been the major bottleneck in deciphering the final bioactive chemical structures of PKs/NRPs by genome mining. RESULTS: Using a large dataset of known chemical structures of macrocyclized PKs/NRPs, we have developed a machine learning (ML) algorithm for distinguishing the correct macrocyclization pattern of PKs/NRPs from the library of all theoretically possible cyclization patterns. Benchmarking of this ML classifier on completely independent datasets has revealed ROC-AUC and PR-AUC values of 0.82 and 0.81, respectively. This cyclization prediction algorithm has been used to develop SBSPKSv3, a genome mining tool for completely automated prediction of macrocyclized structures of NRPs/PKs. SBSPKSv3 has been extensively benchmarked on a dataset of over 100 BGCs with known PKs/NRPs products. AVAILABILITY AND IMPLEMENTATION: The macrocyclization prediction pipeline and all the datasets used in this study are freely available at http://www.nii.ac.in/sbspks3.html. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: In silico prediction of site-specific kinase-substrate relationships (ssKSRs) is crucial for deciphering phosphorylation networks by linking kinomes to phosphoproteomes. However, currently available predictors for ssKSRs give rise to a large number of false-positive results because they use only a short sequence stretch around phosphosite as determinants of kinase specificity and do not consider the biological context of kinase-substrate recognition. RESULTS: Based on the analysis of domain-specific kinase-substrate relationships, we have constructed a domain-level phosphorylation network that implicitly incorporates various contextual factors. It reveals preferential phosphorylation of specific domains by certain kinases. These novel correlations have been implemented in PhosNetConstruct, an automated program for predicting target kinases for a substrate protein. PhosNetConstruct distinguishes cognate kinase-substrate pairs from a large number of non-cognate combinations. Benchmarking on independent datasets using various statistical measures demonstrates the superior performance of PhosNetConstruct over ssKSR-based predictors. AVAILABILITY AND IMPLEMENTATION: PhosNetConstruct is freely available at http://www.nii.ac.in/phosnetconstruct.html.

MOTIVATION: In silico methods are being widely used for identifying substrates for various kinases and deciphering cell signaling networks. However, most of the available phosphorylation site prediction methods use motifs or profiles derived from a known data set of kinase substrates and hence, their applicability is limited to only those kinase families for which experimental substrate data is available. This prompted us to develop a novel multi-scale structure-based approach which does not require training using experimental substrate data. RESULTS: In this work, for the first time, we have used residue-based statistical pair potentials for scoring the binding energy of various substrate peptides in complex with kinases. Extensive benchmarking on Phospho.ELM data set indicate that our method outperforms other structure-based methods and has a prediction accuracy comparable to available sequence-based methods. We also demonstrate that the rank of the true substrate can be further improved, if the high-scoring candidate substrates that are short-listed based on pair potential score, are modeled using all atom forcefield and MM/PBSA approach. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics Online.

BACKGROUND: Enzymes belonging to acyl:CoA synthetase (ACS) superfamily activate wide variety of substrates and play major role in increasing the structural and functional diversity of various secondary metabolites in microbes and plants. However, due to the large sequence divergence within the superfamily, it is difficult to predict their substrate preference by annotation transfer from the closest homolog. Therefore, a large number of ACS sequences present in public databases lack any functional annotation at the level of substrate specificity. Recently, several examples have been reported where the enzymes showing high sequence similarity to luciferases or coumarate:CoA ligases have been surprisingly found to activate fatty acyl substrates in experimental studies. In this work, we have investigated the relationship between the substrate specificity of ACS and their sequence/structural features, and developed a novel computational protocol for in silico assignment of substrate preference. RESULTS: We have used a knowledge-based approach which involves compilation of substrate specificity information for various experimentally characterized ACS and derivation of profile HMMs for each subfamily. These HMM profiles can accurately differentiate probable cognate substrates from non-cognate possibilities with high specificity (Sp) and sensitivity (Sn) (Sn = 0.91-1.0, Sp = 0.96-1.0) values. Using homologous crystal structures, we identified a limited number of contact residues crucial for substrate recognition i.e. specificity determining residues (SDRs). Patterns of SDRs from different subfamilies have been used to derive predictive rules for correlating them to substrate preference. The power of the SDR approach has been demonstrated by correct prediction of substrates for enzymes which show apparently anomalous substrate preference. Furthermore, molecular modeling of the substrates in the active site has been carried out to understand the structural basis of substrate selection. A web based prediction tool http://www.nii.res.in/pred_acs_substr.html has been developed for automated functional classification of ACS enzymes. CONCLUSIONS: We have developed a novel computational protocol for predicting substrate preference for ACS superfamily of enzymes using a limited number of SDRs. Using this approach substrate preference can be assigned to a large number of ACS enzymes present in various genomes. It can potentially help in rational design of novel proteins with altered substrate specificities.

Sequence data arising from an increasing number of partial and complete genome projects is revealing the presence of the polyketide synthase (PKS) family of genes not only in microbes and fungi but also in plants and other eukaryotes. PKSs are huge multifunctional megasynthases that use a variety of biosynthetic paradigms to generate enormously diverse arrays of polyketide products that posses several pharmaceutically important properties. The remarkable conservation of these gene clusters across organisms offers abundant scope for obtaining novel insights into PKS biosynthetic code by computational analysis. We have carried out a comprehensive in silico analysis of modular and iterative gene clusters to test whether chemical structures of the secondary metabolites can be predicted from PKS protein sequences. Here, we report the success of our method and demonstrate the feasibility of deciphering the putative metabolic products of uncharacterized PKS clusters found in newly sequenced genomes. Profile Hidden Markov Model analysis has revealed distinct sequence features that can distinguish modular PKS proteins from their iterative counterparts. For iterative PKS proteins, structural models of iterative ketosynthase (KS) domains have revealed novel correlations between the size of the polyketide products and volume of the active site pocket. Furthermore, we have identified key residues in the substrate binding pocket that control the number of chain extensions in iterative PKSs. For modular PKS proteins, we describe for the first time an automated method based on crucial intermolecular contacts that can distinguish the correct biosynthetic order of substrate channeling from a large number of non-cognate combinatorial possibilities. Taken together, our in silico analysis provides valuable clues for formulating rules for predicting polyketide products of iterative as well as modular PKS clusters. These results have promising potential for discovery of novel natural products by genome mining and rational design of novel natural products.

BACKGROUND: Secondary metabolites biosynthesized by polyketide synthase (PKS) and nonribosomal peptide synthetase (NRPS) family of enzymes constitute several classes of therapeutically important natural products like erythromycin, rapamycin, cyclosporine etc. In view of their relevance for natural product based drug discovery, identification of novel secondary metabolite natural products by genome mining has been an area of active research. A number of different tailoring enzymes catalyze a variety of chemical modifications to the polyketide or nonribosomal peptide backbone of these secondary metabolites to enhance their structural diversity. Therefore, development of powerful bioinformatics methods for identification of these tailoring enzymes and assignment of their substrate specificity is crucial for deciphering novel secondary metabolites by genome mining. RESULTS: In this work, we have carried out a comprehensive bioinformatics analysis of methyltransferase (MT) domains present in multi functional type I PKS and NRPS proteins encoded by PKS/NRPS gene clusters having known secondary metabolite products. Based on the results of this analysis, we have developed a novel knowledge based computational approach for detecting MT domains present in PKS and NRPS megasynthases, delineating their correct boundaries and classifying them as N-MT, C-MT and O-MT using profile HMMs. Analysis of proteins in nr database of NCBI using these class specific profiles has revealed several interesting examples, namely, C-MT domains in NRPS modules, N-MT domains with significant homology to C-MT proteins, and presence of NRPS/PKS MTs in association with other catalytic domains. Our analysis of the chemical structures of the secondary metabolites and their site of methylation suggested that a possible evolutionary basis for the presence of a novel class of N-MT domains with significant homology to C-MT proteins could be the close resemblance of the chemical structures of the acceptor substrates, as in the case of pyochelin and yersiniabactin. These two classes of MTs recognize similar acceptor substrates, but transfer methyl groups to N and C positions on these substrates. CONCLUSION: We have developed a novel knowledge based computational approach for identifying MT domains present in type I PKS and NRPS multifunctional enzymes and predicting their site of methylation. Analysis of nr database using this approach has revealed presence of several novel MT domains. Our analysis has also given interesting insight into the evolutionary basis of the novel substrate specificities of these MT proteins.