Posters - Schedules

Posters Home

Lytic polysaccharide monooxygenase (LPMO) is a copper-dependent redox enzyme and
according to CAZy is classified either as cellulolytic or chitinolytic. According to CAZy, there are eight families of LPMO namely AA9, AA10, AA11, AA13, AA14, AA15, AA16,
and AA17, where AA stands for Auxiliary Activity. Here, the goal is to use existing 3D
structures and AI-based models of the remaining LPMO sequences and train a machine
learning model to identify and classify LPMO as cellulolytic or chitinolytic activity. This was done with the help of extracting structural features such as surface depth, accessible area, electrostatic charge distribution, and geometric features (independent features that define the shape and are not based on enzyme reaction mechanism). Using these we identified the spheric re-entrant face, toric re-entrant face and Solvent excluded surface area features as highly significant with a high signal-to-noise ratio/significance using ensemble feature selection method. The features will be extracted from OBJ and STL format using Pymol, MdTraj, Biopython, Open3d, and Bio3D tools. The model thus trained using structural features will enable identifying and annotating newer LPMO sequences belonging to one of the eight AA families.

This study analyzes mutations in SARS-CoV-2 structural proteins, namely S, N, M, and E. in strains prevalent in Pakistan. A total 467 sequences of each structural proteins from Pakistani strains were retrieved from NCBI along with one Wuhan sequence as reference. Multiple sequence alignments of each structural protein sequences were generated. Mutations were grouped by lineage and classified by variants. DynaMut2 and MutPred used to analyze the mutation’s effect on structural stability and molecular impact, respectively. Each variant's S protein structures was modeled using VMD, while HDOCK and PRODIGY predicted binding energies with ACE2 receptor. From the 467 sequences, 12 were Wildtype, 18 were Alpha, 1 was Beta, 3 were Delta, and 205 were Omicron. The S protein had 142 mutations, followed by N protein with 58, M protein with 16, and E protein with seven. In the spike protein, 15 mutations stabilized the closed state structure, and 14 stabilized the open state structure. Sixteen surface glycoprotein mutations had molecular and phenotypic impacts. Beta and Omicron variants showed the maximum binding affinities with the ACE2 receptor. The study's findings can direct future research on monitoring COVID-19 structural protein mutation properties.

Transglutaminase 2 (TG2) is an enzyme that has important functions in various physiological processes and pathologies, including wound healing, celiac disease and cancer. The aim of this project is to investigate the role of calcium-dependent TG2 in celiac disease using molecular dynamics simulations and coarse-grained modeling. TG2 is involved in the crosslinking of proteins and is known to be overexpressed in the small intestine of celiac disease patients. The activity of TG2 is regulated by the presence of calcium ions, which bind to specific sites on the enzyme. We will use computational methods to study the conformational dynamics of calcium-dependent TG2 and its physiological properties, and explore the effects of non-synonymous single nucleotide polymorphisms (nsSNPs) on the conformational change of TG2 and its association with celiac disease. Specifically, we are looking at four setups of simulation, where the open conformation with and without calcium, and closed conformation with and without calcium are explored. Our hypothesis of calcium-dependent conformational change is explored at long-time scale dynamics in an explicit solvent. The results of this study will contribute to a better understanding of the molecular mechanisms underlying celiac disease and may lead to the development of new therapeutic strategies targeting calcium-dependent TG2.

Polycystic ovary syndrome is a prevalent endocrine illness that affects women of reproductive age. PCOS has been linked to metabolic and reproductive issues like infertility. Natural substances have received interest as prospective therapeutic agents for PCOS .PREGNENOLONE-7,9(11)-DIEN is a natural steroid molecule that has shown promising results in the treatment of PCOS. This chemical has been demonstrated to have anti-inflammatory, antioxidant, and insulin-sensitizing effects. This work performed molecular docking and simulation studies with the human obesity protein (1AX8), which is important in regulating lipid metabolism and energy balance to investigate the potential of PREGNENOLONE-7,9(11)-DIEN as a treatment agent for PCOS. The goal was to study the ligand-protein interaction and assess its potential therapeutic effect. According to molecular docking, ligand binds to the 1AX8 active site with a high affinity and stable contacts, suggesting that it might modulate the activity of this protein. The ligand-protein complex's stability over time was further supported by molecular dynamics simulations, which also revealed information about the complex's interactions and conformational changes. Hence, this research offers insightful information about the molecular interaction between ligand and protein, and it raises the possibility that PREGNENOLONE-7,9(11)-DIEN may be a useful PCOS treatment.

The development of drugs for neurological disorders focuses on voltage-gated sodium channels, with Nav1.7 being one of the nine isoforms involved in neuropathic pain. The recently solved structure of Nav1.7 offers promising pain-related drug candidates. Conotoxins, small peptides from cone snail venom, have diverse compositions and biological functions, making them potential painkillers. This study aims to identify potential binding sites in Nav1.7 for four distinct conotoxins using computational analysis. The structure of Nav1.7 was retrieved from PDB, and ZDOCK was used to dock Nav1.7 with the four cone snail venom peptides. PISA analysis on interacting residues was performed. The results demonstrate that C. geographus, C.ermineus, and C.textile bind primarily to the pore-forming regions of domains II and III, while C. consors interacts with the pore-forming region of domain I. C. consors µ-Conotoxin exhibits distinct hydrophobic interactions in domain I, making it a candidate for a unique Nav1.7 blocker. Further molecular simulation analysis will provide insight into the protein-peptide complex. Utilizing toxins will enhance understanding of channel biophysical and pharmacological characteristics, particularly in distinguishing specific channel activity.

Transcription factors recognize specific DNA sequence motifs but also positions with no motifs for which DNA shape is believed to play an important role. Zentner et al., in 2015, have earlier used chromatin endogenous cleavage coupled with sequencing (CHEC-seq) to try and capture the kinetics of these interaction.They identified that transcription factors identified binding sites at temporally different time scales and can be divided into fast and slow sites based on their time period of occurrence. The fast sites showed consensus sequence motifs while the slow sites did not. It was proposed that the binding specificity for the slow sites were governed by the shape profile of these sites that showed high correlation to the shape of the fast sites. The idea of slow sites and the centrality of shape in their recognition was later disputed (Rossi et al.,2017). We propose to examine the contradictory reports using features of evolutionary conservation, DNA shape and conformational dynamics generated by our tool DynaSeq (Andrabi et al. 2017). We suggest that a high conservation at the sequence level and significant relationship in the shape profiles in these regions would serve as a stronger benchmark and functional relevance of these sites.

The web-based structure viewer iCn3D functions as a platform for comparing (AlphaFold) predicted structures with experimentally determined structures deposited to the Protein Data Bank (PDB), or with other (AlphaFold) predicted structures. A variety of algorithms and methods have been made available, such as VAST (Vector Alignment Search Tool), Foldseek, TM-align, and VAST+. iCn3D also maps annotations onto (AlphaFold) predicted structures, such as conserved domain footprints, functional sites, sites of sequence variation obtained from dbSNP and ClinVar, as well as post-translational modifications. Structure- or sequence-based alignments and annotations are shown via a combination of 3D structure displays, 2D schematic diagrams, and 1D sequence and sequence-tracks displays. Annotations can be exported with Node.js or Python scripts. Users’ work can be saved and shared with others via long-lived short links/URLs or iCn3D PNG images, which reproduce the specific custom views and make them available for further interactive analysis. Users can also view 3D structures in Virtual Reality (VR) using iCn3D. Users can make selections, change style or color, and perform interactoin analysis in VR. The source code of iCn3D can be found at https://github.com/ncbi/icn3d.

The unprecedented outbreak of SARS-CoV-2 required a series of structural studies related to transmembrane serine protease type II (TMPRSS2) due to its importance in viral infection. However, most of them have been hampered by the lack of a crystallized structure of this protein. For this reason, this study aimed to characterize in silico and computationally predict the three-dimensional structure of TMPRSS2 from Mus musculus. The modeling was performed by homology, threading, ab initio strategies, and Alphafold. The best model was chosen according to analysis with ModFold v8.0 and PDBsum. On the other hand, molecular dynamics simulation (MDS) was performed with GROMACS 18.2 to evaluate the conformational stability of the selected model. The binding affinity of the mouse TMPRSS2 complexed with 10 different inhibitors was analyzed by molecular docking. The MDS results showed that the 3D model of TMPRSS2 using SWISS-Model presented higher stability according to RMSD values. Molecular docking revealed that ZINC64606047 has a better binding affinity to the protein. In conclusion, we were able to characterize and obtain a 3D model of the TMPRSS2 protein from Mus musculus and computationally proved its higher affinity for the specific inhibitor ZINC64606047. Financial Support: VRIP-UNMSM and PROCIENCIA-CONCYTEC (Contrato Nº 390-2019-FONDECYT)

A large proportion of proteins are completely or partially embedded in the cell membrane. However, most algorithms assessing the changes in protein structure induced by amino acid substitutions are designed for globular proteins and do not consider the physico-chemical characteristics of the lipid bilayer. We present Missense3D-TM, a program specifically designed to provide a structure-based assessment of the impact of variants occurring in transmembrane regions.
A dataset of 3,346 missense variants (2,197 damaging and 1,149 neutral, in 746 proteins) and 772 3D structures was used for development. Close homologues between the training and testing datasets were removed but a similar pathogenic to benign ratio maintained in both sets.
On the testing set, Missense3D-TM outperformed the standard Missense3D algorithm for globular proteins: sensitivity 58% versus 35%, specificity 81% versus 89%, Mathews correlation coefficient (MCC) 0.37 versus 0.27, accuracy 66% versus 53% (p <1x10-10, two-tailed McNemar’s test). By comparison, the predictor mCSM-membrane achieved 52% sensitivity, 81% specificity, MCC of 0.31 and 61% accuracy (p=0.06).
Missense3D-TM will assist researchers seeking to understand why an engineered or naturally-occurring amino acid substitution occurring in a transmembrane protein might cause changes in protein folding. A web server implementing Missense3D-TM is available at http://missense3d.bc.ic.ac.uk/.

Recent improvements in computational and experimental techniques for obtaining protein structures have resulted in an explosion of 3D coordinate data. For example, while the AlphaFold DB only had 50 gigabytes of protein structure models in mmCIF and PDB formats in year 2021, it hosts 470 times more predicted protein structures (23 terabytes) in 2022. To cope with the ever-increasing sizes of structure databases, this work proposes the Protein Data Compression (PDC) format, which compresses coordinates and temperature factors of full-atomic and Cα-only protein structures. Without loss of precision, PDC compression using delta encoding results in 69% to 78% smaller file sizes than PDB and mmCIF files with standard GZIP compression. It uses approximately 60% less space than existing compression algorithms specific to macromolecular structures, such as BinaryCIF, MMTF and PIC. PDC optionally performs lossy compression with minimal sacrifice of precision by a combination of Cartesian space and torsion space representation. This allows reduction of file sizes by another 79%. Conversion among PDC, mmCIF and PDB formats are typically achieved within 0.02 seconds. The compactness and fast reading/writing speed of PDC makes it valuable for storage and analysis of large quantity of tertiary structural data. PDC is available at https://github.com/kad-ecoli/pdc.

We present pyPept, a set of python scripts to manipulate and analyze peptide molecules using the BILN format to generate atomistic 2D and 3D representations. The scripts allow the analysis of natural and modified peptides that are assembled based on personalized monomer dictionaries. From the line notation, the peptide construct can then be represented as an RDKit object for further prediction of properties and chemical structures.

One important task is the generation of relevant peptide conformers with correct PDB atom naming from the pyPept/RDKit object. We found that including secondary structure (SecStr) restraints during the conformer prediction is necessary. For this end, a similarity-based tool was developed to assign SecStr motifs to the peptides based on a dataset of bioactive conformers available in the PDB. The restraints are included in the distance bound matrix that is subsequently implemented by the ETKDGv3 method from the RDKit. The obtained peptide conformers can embed cyclic restrictions, as well as conserve SecStr elements key for their active conformations, as shown for a few illustrative examples.

In most cases the function of proteins in cellular processes is achieved through interactions with other proteins. Predicting these interaction sites is a fundamental problem. Computational methods for predicting interaction sites have been extensively studied, with structure-based programs providing the highest accuracy. However, the limitation of these programs is the availability of protein structures, which are outnumbered by protein sequences by two orders of magnitude. We introduce Seq-InSite, the first sequence-based model that matches or surpasses structure-based ones. Seq-InSite exceeds the performance of sequence-based models by up to 52%, in terms of area under precision-recall curve, and proves to be superior to state-of-the-art structure-based predictors in all but one test. Seq-InSite achieves this by utilizing the power of MSA-transformer and ProtT5 contextual embedding, within an ensemble of LSTM and MLP architecture. The MSA-transformer captures information from multiple sequence alignments, while ProtT5 addresses the shortcomings of MSA-transformer by using a gargantuan structure and being reliable in calculating proper protein embeddings without the need of an alignment. Seq-InSite is available as a web server at seq-insite.csd.uwo.ca and source code at github.com/lucian-ilie/seq-insite.

Proteins interact through their interfaces, and dysfunction of protein-protein interactions (PPIs) has been associated with various diseases. Therefore, investigating the properties of the drug-modulated PPIs and interface-targeting drugs is critical. Here, we present DiPPI (Drugs in Protein-Protein Interfaces), a two-module website that can be utilized for drug repurposing studies focusing on interface-bound drugs. On the interface module, we extracted properties of interfaces such as hotspots and post-translational modifications of drug-binding residues. On the drug side, we curated a list of drug-like small molecules and FDA-approved drugs from various databases and extracted those that bind to the interfaces. We clustered the drugs based on their molecular fingerprints and provided their drug properties, including Lipinski’s rules. Using this dataset, we docked the HIV protease inhibitors tipranavir and indinavir to the EGFR-ERBB2/HER2 interface and EGFR-ERBB3/HER3 interface, indicating that these drugs can be used to modulate the Ras/Raf/MEK/ERK pathway to suppress metastasis. Our dataset contains 534,203 interfaces for 98,632 proteins, of which 55,135 bind to a drug-like molecule. 2,214 drug-like molecules and 335 FDA-approved drugs are found in the interface region. DiPPI's well-curated and organized interface and drug data offer users an easy-to-follow framework for drug repurposing studies.