ABSTRACT
Purpose of this paper is to study the application of small molecule docking algorithms to the problem of protein-peptide docking. Protein-peptide docking is significant because it can be used to aid in the development of peptide-based drugs, which can provide safe and effective ways to treat many diseases. The present study, which aims to strengthen and extend our knowledge about the strengths and weaknesses of small molecule docking algorithms when it comes to docking peptides, is clearly important and of significant interest. To construct database, the study will retrieve protein sequences from the Protein Data Bank (PDB) with the help of Swiss PDB viewer.
The protein sequences will be retrieved using BLAST algorithm, which allows accurate search of homologous sequences from the PDB. The discusion part is based on known protein and peptide complexes used in deriving preferences. Spatial position specific scoring matrices, used for describing the binding-sites and their environment in globular proteins for each type of amino acid in bounding peptides has been applied. Particular attention is given to systematically docking 53 peptides from the LEADS-PEP database using Vina and PLANTS. Results were analyzed using standard analytical and statistical methods and appropriate conclusions drawn. The analysis of the study carried out by using Vega ZZ to calculate the RMSD results that that the program (VINA) was found successful to find the good peptide pose that binds the protein. And to further find out the best scored RMSD between VINA and Plant indicates that 74% (- negative) results indicating Vina is better than plant. In this regard the researchers should be smart enough to utilize this program in their overall comparison between Vina and Plant for their analysis of the good peptide.
CHAPTER 1: INTRODUCTION
Overview ofthe Study
There is growing interest in the development of novel peptide-based drugs. Thus, there is an essential need for developing, identifying and validating effective methods that could help enable the discovery of novel peptide-based drugs. In particular, there is a great need for the development or validation of effective computational docking methods for use with peptides. Indeed, recent research has focused on validating small molecule docking methods for use with peptides and on developing new, peptide-specific docking methods (Bhachoo et. al, 2017). One study in particular studied the usefulness of several small molecule docking programs for reproducing the x-ray poses of 53 protein-peptide complexes that have been collected in the freely available LEADS-PEP database. According to the study authors, the docking programs Vina and Surflex-dock produced the best results, although there’s plenty of room for progress, especially as it pertains to larger peptide ligands (Ciemny, 2017). The main purpose of the present study is to extend this work and study the usefulness of the proven small molecule docking programs Vina and PLANTS for docking the 53 LEADS-PEP peptides to their corresponding protein targets.
Peptides design and discovery using in vitro appraoches
Peptides are small polymers of amino acids that have been used for many years in the field of biotechnology. For example, they have been used to help produce antibodies. It is becoming increasingly clear that peptides can also be used as pharmaceutical agents (Geng et. al, 2017). In particular, peptides can be developed to modulate clinically important proteins and protein-protein interactions (Geng et. al, 2017). The use of antimicrobial peptides (AMPs) holdsgreat promise, especially in the treatment of bacterial infections (Gordon et al., 2005; Marr et al., 2006). Cell penetrating peptides (CPPs) also hold promise as drug delivery vectors (Gordon et al., 2005; Marr et al., 2006). For example, CPPs can be developed to transport therapeutic compounds as diverse as drugs, SiRNAs or other peptides into target cells.
Notwithstanding the importance of nucleic acids, carbohydrates, or lipids in ligand-target interactions, the effectors of many signal transduction processes are peptides. These can be fragments of proteins or stand-alone hormones, cytokines, toxins, antimicrobials, and many other types of peptides. Truncation of the sequence or non-natural amino acid replacement usually leads to antagonist or inverse agonist derivatives. The nine-residue leptin receptor antagonist Allo-aca follows these design rules and shows opposite effects to the agonist both in vitro and in vivo. Moreover, a library based on the agonist or conformational restriction may allow the selection of peptides with antagonist properties.
PeptiDream uses its Peptide Discovery Platform System (PDPS) technology to generate macrocylic/constrained peptides against multiple targets of interest for major pharmaceutical companies like Merck & Co. The value of the peptide discovery collaboration was not disclosed. The partnership gives Merck rights to develop and commercialize all therapeutic peptides resulting from the collaboration. The pharma giant will also retain an option to nonexclusively license the PDPS technology in the future. The platform applies three core technologies such as “Flexizyme” an in vitro evolved artificial ribozyme (RNA catalyst) which according to PeptiDream can efficiently catalyze the aminoacylation of any non-standard amino acid on to any tRNA. Translation, cyclization, and peptide modifying technology, a flexible cell-free transcription-translation system compatible with Flexizyme-prepared nonstandard aa-tRNA pairs, and designed for efficient incorporation of such nonstandard amino acids without the high rate of mis-incorporation events seen with other cell free translation systems. PD Display, the company’s self-developed in vitro display methodology designed to avoid the drawbacks of conventional phage display, mRNA display, and ribosome display methodologies.
Computational peptide modeling, design and discovery
Several in silico modeling programs, mainly docking methods, exist for the modeling design and identification of small molecule ligands such as Dock, AutoDock, FlexX and similar methods (Ghosh et al., 2006). Indeed, small molecule docking programs have proven to be quite useful in modeling protein-ligand complex structures and in desinging and identifying new small molecule ligands for known targets.Conversely, and despite the promising future of peptides, the use of in silicomodeling and simulation methods, including docking, for peptides remains underdeveloped. However, in what follows we will see that some methods have been developed and tested recently, including the small molecule docking programs Vina and Surflex-dock. The development and validation of an accurate and efficient docking program for peptides can help facilitate the cost and time efficient design of novel peptide ligands and help deliver on the therapeutic promise of peptide drugs.
In silico modeling and design of therapeutic peptides can be divided into two pathways: ligand-based or target-based design. Ligand-based design approaches model and design peptides and small molecules based on ligand information alone. Target-based design approaches also exploit protein target information. Modeling based on the structure of a target obviously requires identifying the target, determining its structure, and identifying the ligand interaction or binding site. Ideally, the structural coordinates of the protein target will be in the form of a co-crystal protein-ligand complex, which can then be used to design new molecules or screen a library of molecules against (Anderson, 2003).
Ligand-based modeling is sometimes a mandatory route, especially when the structure of the target is unavailable. For example, many important receptors are membrane proteins which are notoriously difficult to crystallize. In such cases, all that is available is information about active and/or inactive ligands and ligand-based design provides the only path forward. Very often ligand-based design involves using ligand structural and chemical information to produce a pharmacophore (chemico-spatial fingerprint of the active peptide), followed by virtual screening of conformationally rigid or flexible peptides or small molecules with the goal of identifying some compounds with good QSAR scores (quantitative activity / structure relationship) or predicted functional responses Goede et. al, 2010)
We have seen that experimental methods of designing peptides exist by the in vitroscreening of phage-supplied peptide libraries or by the screening of peptides derived from solid-phase synthesis. As described previously, these methods are time-consuming, costly and produce low hit rates. Hence, there is a need for in silico peptide modeling and design methods. Unfortunately, a small number of in silico peptide modeling methods currently exist. Among these methods, one can speak of the MIMETIC algorithm which was used in the design of a peptide complementary to thrombomodulin with the goal of modulating the function of procarboxypeptidase R (Shimomura et al., 2003). The MIMETIC algorithm was also used to design a peptide complementary to HIV-1 reverse transcriptase segments (Campbell et al., 2002). In the MIMETIC algorithm, peptides are generated in an iterative manner and classified using a score that takes into account physicochemical properties.
Another family of in silico design methods of peptides involves the use of molecular dynamics (MD). MD involves predicting the dynamical motions of a molecule by applying and solving Newton’s laws of motion for the molecular structure, typically when it is under the influence of a molecular mechanics force field like the Amber, Charrmm or Gromos force field (reference). MD has been used, for example, in the design of a human interleukin-6 (IL6) antagonist peptide (hIL-6) (Yang et al., 2005). Unfortunately, MD requires very long calculation times and therefore allows only a very limited number of peptides to be tested. Docking methods can also be used in peptide design.
Unfortunately, most protein-ligand docking methods, which are generally used for small molecule design (Dock (Kuntz et al., 1982), AutoDock (Goodsell et al., 1990)), are poorly adapted to the modeling and design of peptides. As such, many research are working on developing peptide docking approaches. For example, FlexPepDock is a peptide docking algorithm which shows promise (Gunther, 2006).
The PepDesign in silico modeling method allows the rapid design of peptides capable of binding to a target molecule. The successful design of an anti-Aβ29-42 peptide (Aβ peptide fragments are associated with the senile plaques characteristic of Alzheimer’s disease) was performed using this method (Decaffmeyer et al., 2006).These peptides, as well as all their derivatives, have been protected by the filing of a patent (Gunther, 2006) for any therapeutic or diagnostic use of Alzheimer’s disease.This PepDesign method is applicable for the modeling of a peptide targeting a peptide, a protein or any molecule whose three-dimensional structure could be known. It could be used, for example, to design a peptide that would disrupt the interface between two proteins and thus prevent the formation of a protein / protein complex involved in the development of a disease. Used on natural amino acids, it is also suitable for the modeling of peptides composed of transformed (D, non-natural) amino acids. This could allow in the future to test the influence of the required modifications, for example to counter the problems of enzymatic digestion, the recognition and target binding efficiency.
Docking peptides Using VINA and PLANTS as implemented in VegaZZ
Computational docking can be used to predict target-ligand affinity and the associated complex structure (protein-ligand binding mode or pose).Docking flexible ligand molecules to proteins or other macromolecular targets is a significant challenge in molecular modeling, especially as the number of flexible bond torsions increases. Significantly, peptide ligands can have many flexible bond torsions.
Previous research investigated the use of Vina to dock 53 peptide ligands from the LEADS-PEP database to their protein targets. Here, we use Vina, as implemented in the molecular modeling and drug design software package VegaZZ, to try and reproduce their results. We also extended previous research by evaluating the PLANTS small molecule docking program on the same LEADS-PEP database. As in the case of Vina, the version of PLANTS that we used is implemented in the VegaZZ modeling environment.
Statement ofthe Research Question
There is increasing interest in using docking to design and develop peptide drug candidates. Previous work studying the utility of small molecule docking algorithms to reproduce the 53 protein-peptide complex structures of the LEADS-PEP database provides a nice starting point for continued research on the protein-peptide docking problem. In the present study, we used the LEADS-PEP database and Vina (as implemented in VegaZZ) to try and reproduce the results of previous studies. We also extended previous work by testing PLANTS using the LEADS-PEP database.
Goals and Objectives
Goals
The main goal of the present study is to study the application of small molecule docking algorithms to the problem of protein-peptide docking. Particular attention is given to systematically docking 53 peptides from the LEADS-PEP database using Vina and PLANTS. Results were analyzed using standard analytical and statistical methods and appraoriate conclusions drawn.
Objectives
1. Obtain and prepare the LEADS-PEP database
2. Use Vina, as implemnted in VegaZ, to dock the 53 LEADS-PEP peptide ligands into the binding sites of their protein targets
3. Analyze the Vina docking results by comparing the structural coordinates of the docking pose predictions with the corresponding x-ray protein-peptide complex structural coordinates
4. Repeat 1-3 using PLANTS
5. Analyze and interpret all docking resultsand draw appropriate conclusions
Significance of the Study
The discovery of a new drugis a complex process which can take years and huge sums of money and skill to successfully complete. Audie and Swanson hold that biological screening approaches – which involve tedious, complex, and cumbersome processes – have been essentially the only approach used to date in the identification of novel protein-peptide interactions (Lee et. al, 2017). Audie and Swanson further explain that protein-peptide docking is an in silico approach for identifying and predicting protein-peptide interactions and structures. In this view, protein-peptide docking can be used as a high throughput approach of designing and developing peptide-based drugs.
The evolution of technology requires the application of high throughput techniques in the development of novel drugs, including protein-peptide based drugs. According to Hauser and Windshugel, the development of peptide-based drugs comprise over 10% of the drugs approved in the recent past and the form the best drugs such as boceprevir and telaprevir used in the treatment of hepatitis C virus (188). Currently, many peptides drugs have been approved and thousands are under active investigations.
Protein-peptide docking is significant because it can be used to aid in the development of peptide-based drugs, which can provide safe and effective ways to treat many diseases. Rentzsch and Renard assert that protein-peptide docking provides a promising way of optimizing pharmaceutical peptides. Therefore, the present study, which aims to strengthen and extend our knowledge about the strengths and weaknesses of small molecule docking algorithms when it comes to docking peptides, is clearly important and of significant interest.
CHAPTER 2: EXPERIMENTAL AND METHODS
2.1 EXPERIMENTAL
Peptides and Peptide Derivatives
The ideal drug is almost a myth: it is a molecule of low molecular weight, soluble, orally bio-available, capable of chemically and almost perfectly mimicking a natural ligand and without having bad side effects. Moreover, its production should be easy and inexpensive. Finally, its administration to the patient should be as easy as possible.The chemistry of the drug has always been based on the exploitation and transformation of natural medicinal substances, in particular products derived from plants, thereby copying our ancestors’ use of them. This source is exhausted. We want to prove the exhaustion of the pipelines of the pharmaceutical laboratories described by the American financial analysts like the pipeline problem. This is reflected in the fact that many operating patents arrive at the end of their lifetime as more and more new molecules are introduced to the Food and Drug Administration (Surowiecki, 2004).
A very recent report by the American Congress indicates that the ratio between the number of active new molecules deposited and the R & D budgets of firms has decreased significantly in more than 20 years (Howlett, 2006). This indicates that innovation in this area is increasingly difficult and costly. Indeed, the development of a new therapeutic molecule costs between 200 and 800 million euros (Anon., 2003; Goozner, 2004) and takes about 10 years.An increasingly envisaged way by pharmaceutical laboratories to propose original therapeutic alternatives is to design drugs of the peptide or peptidomimetic type.
Therapeutic peptides can be classified into three categories (Sato et al., 2006): first, natural peptides, also called bioactive peptides that are either produced by the body or derived from proteins (Watt, 2006); Secondly, peptides derived from genetic libraries by phage display (Sergeeva et al., 2006); And finally, peptides derived from chemical libraries produced by solid phase synthesis (Shin et al., 2005). There may also be mentioned vector peptides which are not, strictly speaking, drug molecules but which serve to transport therapeutic molecules (Mae et al., 2006). Among these transport peptides are, for example, Penetratin derived from the Antennapediahomeodomain, designed for the delivery of bioactive proteins (Derossi et al., 1994) or Transportan, a chimeric peptide derived from Galanin and Mastoparan, designed to transport interfering RNAs (Pooga et al., 1998), and the like. Finally, a last research pathway using peptides is that which uses them not as finished products but as tools to design peptidomimetics, small molecules that mimic the bioactive properties of the peptide (Patch et al., 2002).
The discovery of the first bioactive peptide can be dated to more than 80 years. It was in 1923 that Banting and Macleod were awarded the Nobel Prize for Medicine for their discovery of insulin whose role in the treatment of diabetes is known. During this period, and up to the 1960s, the synthesis of peptides by conventional chemical processes took up to several months, rendering impossible any industrial exploitation. It was not until 1963, when Merrifield developed the method of solid peptide synthesis (Merrifield, 1963) for which he was awarded the Nobel Prize in Chemistry in 1984, and also thanks to the development (HPLC) (Chen et al., 1995) or SPE (solid phase extraction) (Kamysz et al., 2004), that the industrial use of Peptides not originating from the natural environment has become possible. On the other hand, over the past twenty years, the development of the phage display technique (Smith, 1985) and progress in the control of this method have led to the production of high performance peptide libraries leading to the selection of Peptides (Cortese et al., 1996).
Since these technological breakthroughs, peptides, like proteins, are considered to be therapeutic molecules of the future. In 2004, more than 20% of the medicines in the top 200 were protein or peptide-based, with sales reaching $ 40 billion (McGee, 2005), or about 10% of the total ‘pharmaceutical industry. In 2004, between 600 and 700 peptides were in developmental stage and more than 150 were at different clinical stages (McGee, 2005). They are found in various fields such as the treatment of certain forms of cancer, AIDS, osteoporosis, neurodegenerative diseases (Lien et al., 2003).
Peptides offer several advantages over the small molecules that make up the traditional drugs. The first advantage is that, often representing the smallest functional part of a protein, they offer efficiency, selectivity and specificity that small molecule designers find difficult to achieve (Hummel et al., 2006). An example of this is the recent work on NAP, an 8-residue peptide, the smallest Active Dependent Neuroprotective Protein (DPA) fragment showing neuroprotective activity (Gozes et al., 2006). This peptide is currently in clinical trials.
A second advantage, and not least, is that the peptide degradation products are amino acids, which greatly eliminates the risks of toxicity (Loffet, 2002).
Thirdly, peptides accumulate little in tissues because of their short half-life, which, it is true, is also a disadvantage when it comes to duration of action.The design of peptides for therapeutic use, however, remains extremely complicated and raises a number of challenges. First, the cost of production is higher than that of small molecules and the quantities needed are greater (Marx, 2005), mainly when the route of administration to the patient is not oral. Secondly, the peptides do not obey the famous rules of the five (Lipinski et al., 2001), especially regarding their molecular weight. According to these rules, peptides would be unable to pass from the digestive tract to the circulatory system. Therefore, if this criterion is used conventionally in the pharmaceutical industry, the peptides are considered to be little absorbable orally. For this reason, the administration of the therapeutic peptides is often intravenous, which inevitably causes a lack of comfort for the patient.
However, alternative delivery strategies exist, such as mucosal administration (Prego et al, 2005), nasal (Maggio, 2005) or pulmonary (Skyler et al., 2001), for example. Other groups have developed coatings allowing delayed release of the peptide and therefore the spacing of the injections. This is the case of the company Roche which has put on the market the Pegasys TM which is a pegylated formulation of Interferon which makes it possible to limit substantially the number of injections. Still others, such as Bioject, have developed high-pressure gas propellants that allow the peptide to cross the skin without requiring an injection (Stout et al., 2007). Finally, the eligen R technology developed by Emisphere allows the peptide to be coated to allow a transfer from the digestive tract to the blood, allowing oral intake (Malkov et al., 2005). Finally, strategies using carrier peptides are also in full development.
The third aspect to be addressed here is related to the biodegradability of the peptides due to the peptidases of the gastrointestinal, kidney, liver and serum systems. This degradation, which results in a short half-life, of the order of a few minutes or even at best of a few hours, can be countered in various ways. One of the most widely used pathways is the use of modification of the terminal C and N parts (by N-acetylation and N-amidation, for example) of the peptides to prevent the action of exopeptidases. Also considered is the use of D or non-natural amino acids, which are less subject to the action of these enzymes. Cyclizations, whether by disulfide bridges, lactam ring formation or Nter-Cter bonds, are also used to counteract the action of proteases (Werle et al., 2006).
Finally, Pegylation, that is to say the binding of poly (ethylene glycol) units, is an increasingly used strategy. Some groups have also shown that it is possible to retain the function of a peptide while it is partially composed of D amino acids or even that it is possible to model a peptide Of function given with non-natural amino acids (Lins et al., 2006).These transformations lead to the design of molecules that are less and less related to peptides and more and more to small molecules that bear the essential chemical functions highlighted by the use of peptides. We will speak of peptidomimetics or pseudo-peptides. For these products, the peptide is not an end product but a design model. Conventionally, the design of a peptidomimetic takes place in several steps:
1. Identifying the target protein
2. Identification of the effector (ligand, bioactive peptide or partner protein)
3. Search for the shortest bioactive segment
4. Use of structurally constrained peptide analogues by methylating Cα (Sagan et al., 2001) or Cβ (Birney et al., 1995), or even by cyclizing it
5. Establishment of a pharmacophore
6. Screening of small molecule libraries respecting the chemico-spatial properties of the pharmacophore (Hummel et al., 2006).
Structural Aspect
We know that the structure of the peptide is crucial, in particular because it must imperatively correspond to an active conformation in the presence of the target. On the other hand, it seems essential to be able to verify whether the transformations (modification of the peptide backbone or of the side chains) will leave the pseudo-peptide the possibility of adopting the conformation required for its function.To date, several experimental methods for determining the structure of peptides exist, including circular dichroism (DC) (Holzwarth et al., 1965), Fourrier Transform Infrared Spectroscopy (FT-IR) (Goormaghtigh Et al., 1990), X-ray crystallography or nuclear magnetic resonance (NMR) (Bechinger et al., 1999). These techniques are often cumbersome to implement and are also confronted with various limitations, for example crystallization, aggregation, etc.
On the other hand, the reality is that the data relating to the structure of peptides provided from these different techniques often diverge.This observation indicates that the idea that a sequence gives rise to a unique structure in a given solvent, if it is coherent for most proteins, does not necessarily apply to peptides. This is in particular due to the fact that in comparison with proteins, the peptides have a smaller number of possibilities for intramolecular stabilization and also that a single peptide sequence can lead to several energy solutions that are comparable and therefore probable.If different experimental methods of determining peptide structures exist, to our knowledge only three methods of in silico determination are available to date. The first method of in silico determination of structure is called ROSETTA and was developed in the late 1990s by an American team (Simons et al., 1997).
It is based on the mixed use of homology modeling with known protein segment patterns for local conformation determination, and de novo prediction using a mean field-based form field (MFP). This method gives ten structures, the most probable of which is number 10. The Robetta server using the Rosetta procedure is accessible on the net (Chivian et al., 2003). ROSETTA has been used extensively for the prediction of the three-dimensional structure of proteins (Bonneau et al., 2002, Bradley et al., 2003), and this has been very successful in various editions of the Critical Assessment of Techniques for protein Structure Prediction). It has also been successfully used to construct de novo proteins capable of adopting a pre-selected fold, the best known case being that of Top7, a 93-residue protein comprising two α and five β-strands which has been modeled by The David Baker team (Kuhlman et al., 2003).
Unfortunately, although this method has proven to be useful for proteins, it is not suitable for peptides, probably because of the force field used, based on MFPs, ie a statistical analysis of the structures Of proteins. This field of force tends to fold the peptides like proteins and leads to globular structures (hydrophobic collapse) that do not agree with the data obtained experimentally (Thomas et al., 2006). In any case, the Robetta server does not allow prediction on sequences of less than 20 residues, which clearly indicates that the method is more intended for the study of proteins than peptides.
The second known method for ab initio determination of the peptide structure was previously developed by an Indian group. This approach, called PEPSTR, is based initially on the prediction of the residual secondary structure by the BETATURNS method (Kaur et al., 2004). Then, standard ang angles are assigned to the residuals before an energy minimization using the Amber6 force field. The method makes it possible to treat peptides of 7 to 25 residues. It gives a unique structure.
The last method was developed by Biosiris SA in collaboration with R. Brasseur’s team in Gembloux (Thomas et al., 2006). This method, called PepLook, is specially devoted to the in silico prediction of peptides and is not yet adapted to the prediction of protein structures. It is based on the exploration of the conformational space of the peptide using a set of 64 pairs of angles ΦΨ which allows the reconstruction of any protein of known structure (Etchebest et al., 2005). An iterative stochastic procedure such as that used in many scientific fields (Glick et al., 2002) is used and is further modulated by a Bolzmann-type approach which allows, step by step, to modulate the probability of use Of the different pairs of angles Φ / Ψ for each position as a function of their contribution to the occurrence of good or bad energy structure. This method gives the most stable structure from an energy point of view, called the Prime, but also 98 other low energy structures.
This method is original in that, besides giving the most stable structure, it gives access to parameters on the stability of the peptide, on its propensity to bind external partners, as well as on its polymorphism (or disorder or Structural diversity). Indeed, a stability score is established by comparing the Mean Force Potential (MFP) of each Premium residue with reference values calculated from a non-redundant bank of nearly 500 folded protein structures (Lovell et al .), Which also makes it possible to identify the amino acids most likely to create extramolecular interactions and thus to seek partners in order to increase their stability. Finally, a polymorphism (or disorder) score is also established by the calculation along the sequence of a RMSd [9] (Root Mean Square deviation on a nine residue window) of each of the 98 best energy structures by To the Premium.
The latter is far from being anecdotal if one is to conceive that a given peptide must necessarily be able to adapt its structure to fulfill several different tasks. If one takes the case of a peptide capable of passing a plasma membrane to fulfill its function, it must simultaneously be soluble in a hydrophilic medium (the blood and the cytoplasm), in a hydrophobic medium (the membrane) Organize it in a third structure in the presence of its target. It can present adequate solubility in different media only if its apparent hydrophobicity changes, and therefore if its structure changes. Polymorphism is therefore more than likely a determining factor for the biological functionality of peptides, and as such, it seems important to have access to this information.
This hypothesis has been proposed for vector peptides. The functional vector peptides known to date have common characteristics such as a maximum size of 30 residues, a certain amphipaticity and a positive net charge. However, their mode of operation, and more particularly how they are able to pass through biological membranes, even if it has been extensively studied by biophysical methods (Fischer et al., 2005), has not yet been elucidated. If it is established that the internalization of the peptides is conditioned by their amphiphilicity and the loads they carry, it is also largely dependent on their structure (Deshayes et al., 2005). PepLook was recently used to study the properties of two of these peptides and their mutants. This work clearly shows that structural polymorphism is an essential characteristic for cell penetration since it allows the peptide to adapt its conformation to its environment (Deshayes et al., 2008).
This observation is not insignificant and could be the basis of new methods of design of vector peptides.However, the same method was also used in another domain to predict the organization of the membrane segment of human DGKε (D19-Q23), a monotopic membrane protein of unknown structure and involved in the transformation of diacylglycerol into phosphatidic acid. Finally, in the near future, the method will be used to verify the effect of chemical modifications on the structure of the peptides.
Peptide Nano-Medications
Accurate targeting of protein and peptide (P / P) drugs is essential for their efficacy and pharmacological safety. European scientists have explored the potential use of nanoparticles to deliver these drugs to targeted tissues and organs. Protein or peptide-based drugs have evolved in recent years into very good drugs targeting certain diseases. Due to the highly diversified structure and broad biological activity of the peptides and proteins, these P / P drugs can take the form of hormones, neurotransmitters, structural proteins or even metabolic modulators with an important role as Therapeutic molecules and biomarkers. The Nanobiopharmaceutics (Nanoscale functionalities for targeted drug delivery of biopharmaceuticals) project focused on the production, testing and implementation of many nanoparticle multi-carrier systems for P / P drugs. Thus, NanoToolbox, combined with peptides, has been tested in vitro and in vivo in terms of efficacy for oral or nasal drug delivery or for crossing the blood-brain barrier (BBB).
Partners have generated systems to obtain targeted administration of P / P drugs and test their toxicity, immunogenicity and degradation properties. In addition, they developed P / P polymers and studied their interaction with the cells. The Nanobiopharmaceutics project has established innovative multidisciplinary approaches to the design and synthesis of micro- and nanometric molecular functions for the targeted delivery of therapeutic peptides and proteins. The generated systems were a breakthrough in the field and should result in many clinical applications for the treatment of diseases.
Virtual Screening
Virtual screening of databank libraries has become a method Computer-assisted routine used to identify ligands for targets of therapeutic interest. It should be borne in mind that this technology is very sensitive to 3-D coordinates of the target and generates many false negatives, which can be up to 30%. Also important to producing success is library preparation. Rather than focus on the hit rate, it is more interesting to consider the number of new chemotypes discovered in ligands identified and validated by screening. From this point of view, virtual screening is a natural complement of the medicinal chemist in order to propose molecular chassis capable of rapidly leading to high value-added chemotheques.The selection of scored keys among the first 5% of different functions allows To enrich the final selection in real positives ( Charifson et al., 1999, Bissantz et al. 2000 ).
This method has the advantage of adjusting a screening strategy as a function of Known experimental data. Simply prepare a test library or a small Number of real assets (of the order of a dozen for example) and to mix it with a large Number of molecules supposedly inactive (one thousand for example), to anchor the library with various docking tools, and to rescore the poses obtained with different functions of Scoring. The systematic analysis of enrichments in real positives is done by calculating the Number of real assets in various selection lists determined by simple scoring or multiple. The screening strategy (docking / scoring) giving the best Then be applied to full-size screening. Despite these advantages, this technique cannot be applied in the absence of experimental data (knowledge of several realities chemically diverse assets). In this case, it is necessary to implement more general strategies to eliminate false positives: detection of under-buried ligands ( Stahl and Böhm, 1998 ); Refinement by energy minimization of docking poses ( Taylor et al., 2003 ); Consensus docking by various tools ( Paul and Rognan, 2002 ); Docking on conformations Multiple of the target ( Vigers and Rizzi, 2004 ); Rescoring of multiple poses ( Kontoyianni et al., 2004 ). For the most part, these approaches are quite complicated to implement and Guarantee a wide application to various screening projects.
Computational Methods in Drug Design
The ability to store, retrieve and analyze an enormous amount of information makes the precious computers, and in some cases essential, in biological research (bioinformatics), chemical (chemoinformatics) and consequently in the pharmaceutical industry. We have already mentioned the use of computers and of computer science in the identification of the proteome proteins, in predicting the metabolic stability of a molecule, in the analysis of virtual product libraries (v. Bioinformatics). These applications take advantage of the capabilities of modern computers, even of small size, with appropriate algorithms to analyze huge collections of data (data base) to identify structures, in order to deduce a wide range of properties and give an immediate graphical representation and easily understandable. These are applications extremely useful but not directly related to the design of the drug, as is the case for the molecular modeling (molecular modeling; v. Gubernator and Bohm, 1998).
The latter is a technique that allows to determine the geometry of a molecule and the possible conformations, by calculating for each of them the energy and therefore the probability of existence, the molecular orbital and the electron density, the lipophilicity, the volume, the surface accessible to the solvent. You can simulate the dynamic behavior of the molecule in a vacuum, in a solvent or when interacting with another molecule of biological interest. If the site of interaction is not known if it can evaluate the mode and take into consideration the structural changes necessary to increase the interaction energy; if the site of interaction is not known, they can be processed models that help you understand what it is, and then to design more related molecules based on structural features of other molecules able to positively interact with that site.
The molecular modeling programs are numerous, but essentially are based on two types of approach: that of the quantum and molecular mechanics. The quantum-type programs calculate the properties of the molecule by solving the Schroedinger equation in an approximate, with so-called semi-empirical methods, or more rigorously, with ab initio methods, but generally require more computing power. The molecular mechanics programs steric calculate the energy of a molecule using empirical force fields (force field) formed by the potential functions derived from classical mechanics, as for example the law of Hook.
Another approach widely used for the study of molecular dynamics is that of the simulated quenching (simulated annealing); in this case the molecule in question is described as a dynamic structure where the atomic coordinates are changed with respect to time according to the kinetic energy of the atoms and the forces exerted on them by the surrounding atoms and by the force field in which they are immersed. The information collected through these methods, particularly those on the lowest energy conformation, are normally used for drug design, as you will see later.
Another aspect of computational chemistry which finds particular use in the study and drug design is the determination of quantitative relationships between chemical structure and biological activity (QSAR, Quantitative Structure-Activity Relationships). The method that has had the most successful among many tested is to Hansch (v. Kubinyi, 1993), in which these relationships are described by an equation multiparametric that relates the biological action with chemical-physical properties such as lipophilicity , the electronic distribution and steric structure, defined by suitable parameters. When the equation assumes statistical significance, the variable coefficients provide valuable information on the characteristics needed to produce the biological activity of its series of test drugs, information that can serve both to explain the mechanism of action of the drug class studied, both to program the synthesis of other molecules with the best characteristics. The combined use of QSAR and molecular modeling led to 3D-QSAR (QSAR three-dimensional), which improves the quality and quantity of information obtainable with the traditional technique.
Among the various methodologies developed, the most widespread is the comparative analysis of molecular fields (CoMFA, Comparative Molecular Field Analysis), which is based on the assumption that the molecular interactions of a biologically active substance with their biological target can be described by means of electronic force fields, steric and hydrophobic, and that the molecules can therefore be characterized on the basis of the value of these interactions calculated at a distance and through appropriate molecular probes. The CoMFA analysis results are shown through the coefficient contour maps (electrostatic, steric, hydrophobic), which express the field strength at a particular point in space.
Synthesis and Release ofGrowth Factors Summarized Peptides inSolid Phase forTissue Engineering
In recent years much attention has been paid to the peptides as novel therapeutic agents, since they are flexible to adopt and mimic the local structural characteristics of growth factors. Versatile features to perform synthetic organic manipulations are another feature of the peptides compared to protein-based drugs, such as antibodies. The sustained release of the peptides can be established through different retention mechanisms such as non-covalent physical entrapment, absorption, or covalently using the coupling agents. Considering these mechanisms, our research activity is based on the controlled release of peptides mimicking growth factors (e.g. Peptide derived from BMP-2 human) by porous biomimetic scaffolds and injectable materials. In addition, the controlled release is also obtained from dendrimersbicompetenti able to incorporate and release some components (peptides, drugs, bioactive molecules, cells) in a specific site of the body for a fair period of time. The synthesis of the peptides and dendrimers is carried out using a microwave synthesizer using Fmoc chemistry in the solid phase. The microwave heating enables direct heating “in core” of the reaction mixture, which results in a more uniform and fast heating.
An example of our work is the development of a scaffold polymer or organic-inorganic hybrid material for the sustained release of the peptide of the human BMP-2, already applied in the clinic in the regeneration of bone tissue due to its osteoinductive properties. The strategy involves a double bioactivation of the scaffold; complexationpolyanionic was used to produce some stable interaction, in which the release from a biodegradable polymeric carrier may take place by electrostatic interaction. While, the chemical immobilization of the peptide in the scaffold allows a long-term release. However, since the biodegradation of the support matrix would be the most likely mechanism of release, it is possible to control the kinetics of release of the drug by adjusting the polymer degradation rate. This dual approach is able to ensure appropriate biological response from human mesenchymal stem cells for the regeneration of bone tissue.
2.2 METHODS
To construct database, the study will retrieve protein sequences from the Protein Data Bank (PDB) with the help of Swiss PDB viewer. The protein sequences will be retrieved using BLAST algorithm, which allows accurate search of homologous sequences from the PDB. The homologous sequences will be downloaded and analyzed using PROCHECK to differentiate between protein chains and peptide sequences and isolate them according to their stereo chemical quality. CD-CHIT will be used in clustering proteins based on their similarity scores and sampled for further analysis.
Information aboutSWISS PDBSoftware
The first column shows each of the tools in their normal, safe mode. There is also a link to the .PDB file about the tool (protein database ) that describes its geometry (there are several graphic viewers of .PDB files) The Swiss-PDB viewer works on Mac, Windows, and Linux). An “L” in the PDB file name indicates that the tool is loaded, a “U” indicates that the tool is idle. The energy of the tools (see the note on the energy units ) is also displayed in this column.
The second column shows the potential failures (only if they have been updated by the simulation). There is also a link to the .PDB files in this configuration that results in a failure. The number and the letter following the L or the U informs us about the distorted structure and the set of starting data of the chosen mode. The energy of the mode is also shown. If no potential failure has been discovered to date, the second column contains links to the original files and results (described below).
The third column shows some deformed tool structures, with the index number of the structure, which minimize potential failures. The “MD run” line, for example, “MD run: 3500 K (0.04 RMSD)”, shows the ambient temperature of the simulation required to produce deformations of the tool with the indicated RMSD.Below each potential mode of failure is the “notes” line with additional information and explanations about the mode.The last column of each tool contains links to the starting folders and the results. The letter after the L or U indicates which ambient temperature of the simulation was used to deform the tool. The following table gives the correspondence between the letters and the temperatures:
Z 1500K Y 2000K X 2500K W 3000K V 3500K U 4000K T 4500K
Logic High Temperature / Low Level
If these tool tips were to be treated as organic molecules, the identification of the default structures would correspond to the identification of the reaction processes between the starting structures and the final structures. Based on these structures, one would identify an organic “mechanism” that explains the rearrangement that a structure must undergo to pass from its original form to its final form. In order to do this correctly, a series of geometric optimizations should be calculated on the reagent (starting structure), on the product (final structure) and on one or more transition structures throughout the reactions defined by the mechanism. On the other hand, the high temperatures and the theoretical (relatively) low levels that are used by default in this analysis for the purpose of “crude forcing ” the identification of default structures, provide the final geometric structures. These may correspond to feasible minimal forms of energy and can thus be used as basic structures for developing plausible mechanisms.
The low levels of the theory (RHF / 3-21G / STO-3G) are used for these studies because we are looking for strongly (covalent) tool structures. The forces of the covalent links are such that even at the low level of theory, it is possible to predict correctly their connectivity. The energies of these structures will, however, differ from the levels of the theory. Modeling candidates with a reasonable minimum structure for default calculations requires far more computing resources than optimizations at the highest geometric levels of the structures identified by default because, for example, the 5000 starting structures calculated with GROMACS will be able to optimize geometrically Only 15 tools, this is a consequence of not actively directing the generation of default structures.
Energy Unit
The standard unit used in quantum chemistry to measure the total energy of molecules is the “Hartree”. The basis in quantum chemistry is that a system (“infinite separation”) with a completely dissociated nucleus and electrons corresponds to zero energy (0 Hartree). This means that the separation of a molecule and its constituent atoms and the separation of each atom with its electrons and its individual nucleus (the different protons and neutrons of the nuclei are not included in this separation). As the electrons and nuclei are brought closer together to form the molecule, energy, by definition, decreases (negative value).The only appropriate way to compare energies for calculations in quantum chemistry is to compare the same number of atoms to the same level of the theory. Therefore, each tool and its default structures are considered as separate groups without overlapping between the various basic tools (called structures).
There are two types of minimum energy structure considered in these (all) calculations. The first is the global minimum, the structure that refers to the lowest possible energy for a given number of atoms at a given level of the theory. The other type of structure is the local minimum, a structure that is stable (in quantum chemistry by performing vibratory analyzes on the molecule) and will not spontaneously rearrange itself in a more stable form but with a higher energy than that of the global minimum (basically, of all other stable forms). In theory (although part of the Q-SMAKAS project is to determine the validity of this declaration of principle), a “ready” tool corresponds to the overall minimum energy of these structures and each default structure is a local minimum.
VEGAa ZZ Software and PLANTS Docking
VegaZZ is a computational drug design laboratory. It includes the well known small molecule docking programs PLANTS and Vina. Essentially, PLANTS is a protein-ligand docking algorithm that applies ant colony optimization (ACO), which is a set of stochastic optimization algorithms. PLANTS work by employing artificial ant colony in ascertaining the minimum energy required to form optimal conformation of protein-ligand interaction at the binding site. The docking algorithm imitates the behavior of ants, which establish the shortest path towards their food, and thus, ascertain the conformations with the lowest energy protein-ligand interactions (Chen 81). The use of artificial colony gives trail of artificial pheromone, which the iteration process of docking uses in producing low-energy conformations.
In the docking process, the default parameters and customized parameters will be used to optimized protein-peptide docking process and give accurate determination of the interaction between proteins and peptides. Lamarckian genetic algorithm with the parameters of high accuracy or standard accuracy will be used in compiling docked proteins and peptides (Hauser and Windshugel 189). Protein database (RCSB PDB) and chemical database will provide diverse peptides and proteins that the study will use in creating database and performing protein-peptide docking.
Data Collection
The study will compare docking results of PLANTS and VINA to establish which one is successful between the two. In comparing docking results, the study will use LEADS-PEP protein-peptide benchmark data set. The study will access the data set from the website (leads-x.org) and use it in comparing the docking results. The data set provides physicochemical properties such as peptide length, sequence, heavy atoms, bonds, ring counts, hydrogen-bond acceptor, hydrogen-bond-donor, molecular weight, and probability criterion (Hauser and Windshugel 190).
The study will interpret the outcome of docking using predicted binding energies of protein-peptide interactions. Moreover, the study will interpret the efficacy and safety of the predicted peptide-based drugs using their respective biochemical and chemical proteins as ascertained from ADMET SAR.
The materials to use for the research are;
I. Vega ZZ Computer Program (Vina and PLANTS)
II. Protein Database (RCSB PDB)
III. Lead-pep data base
The materials and method were used to calculate the RMSD for the protein-peptide affinity and to see how much they are able to bind, then to compare between the two docking programs VINA and PLANTS. Also to compare them with the results in the main article.
In the second stage after using lead-pep proteins and peptides data Bank PDB base to re-dock them by using VEGAZZ software. It Involves opening of the protein by using Vega ZZ software. The stage involves a set of activities that start with the removal of the water molecules and running the receptor docking using VINA with protein. After that we have open the inhibitor and run the ligand docking using VINA.
In this study we run VINA docking with the receptor and ligand using Vega ZZ Computer Program. Next is to open the peptide and run the ligand docking using VINA and PLANT along withrunningthe VINA docking with the receptor and ligand using VINA docking research tool. For this we first open the peptide in vegazz and click on file-run script-docking-VINA -ligand-run and open the receptor in Vega ZZ and then click on file-run script-docking-vina-receptor. C-run. On the other hand we had run the reference pep in Vega ZZ and go to the center of the peptide using right click to indicate the center coordinates by pointing on the center and choose chain-change-atom and there is a table appeared that includes the exposed coordinate , it is usually on CA (alpha carbon). These coordinates are then used in the docking calculation.
Data Analysis
Then docking process is started by using the peptide and receptor using Vega ZZ software obtained from the reference. In next stage after the docking has done we will obtain the best 9 peptides poses that are fitting the receptor and the closest pose to the reference. We open the pep-out in Vega ZZ and save each one the 9 of the as mol2 file, do the same with reference and rename it as peptide-out 0. Then we return to the database file in order to save all the 9 best peptides with the reference and merge them together. Next, we open the merged file in Vega ZZ to calculate the RMSD.We have found the 9 of the protein poses compared to the reference when we see the RMSD between 2.00A and 2.50A this will be good result , between 1—2 is better, less than 1 is perfect. If the result is more than 2.50 A, it is unsuccessful and that means the program (VINA) was not successful to find the good peptide pose that binds the protein that will be further discussed in data analysis of the study.
Limitation of the Study
The research involving the use of protein-peptide docking for drug designing is limited in majority of the published applications of docking, the third step is taken widely and the modes of binding ligands have been reproduced. However in all those cases, there is a predetermined binding site and thus the search is limited to that particular region of the selected protein in the docking simulations. The results convinced of such studies indicated that the possibility for applying the same methodology for a larger space than that of a single binding site.
Chapter 3: RESULT AND DISCUSSION
The methodology applied is of high benefit to researchers investigating the structural basis of proteins and protein interactions. It can be used in structures known to bind a peptide. The significance of the method is the ability to provide informative inferences about the site of interaction. It is most appropriate as a readily available method for suggesting further experiments test the interaction. Despite challenges because of lacking data and information, it is currently difficult to modify residues from being studied. However, the expected steady growth in structures has the possibility of guaranteeing additional residues to be considered with time. The new structures are expected to improve each residue profile, as well as the various approaches, provide suitable references to deduce results. Additional data is expected to permit more sensitive residue pairs, which we expect that they will increase the performance. We are determined to develop various modifications to account for the limitations encountered in the experiment, for example, the only special that peptides bind via beta-sheet augmentation.
3.1 RESULTS
The study will compare docking results of PLANTS and VINA to establish which one is successful between the two. In comparing docking results, the study will use LEADS-PEP benchmark data set. The study will access the data set from the website (leads-x.org) and use it in comparing the docking results. The data set provides physicochemical properties such as peptide length, sequence, heavy atoms, bonds, ring counts, hydrogen-bond acceptor, hydrogen-bond-donor, molecular weight, and probability criterion (Hauser and Windshugel 190). Moreover, the study also will predict chemical properties of the drug candidates using chemical database (ChemSpider).
The analysis of the study has been carried out by using Vega ZZ to calculate the RMSD. We will find the nine of the protein poses compared to the reference when we see the RMSD between 2.00A AND 2.50A this will be good result, between 1—2 is better , less than 1 is perfect. If the result is more than 2.50 A it is unsuccessful and that means the program (VINA) was not successful to find the good peptide pose that binds the protein. For this first we convert the receptor and ligand from pdb from pdbqt to mol2 and then we have performed docking for the plant by taking the protein as receptor and the peptide as ligand. After that we have mentioned then coordinating number to run. After that we get the result at the 8 excel file. The data is categorize on the two data files which shows the docked-ligand mol2 and veg azz. Then we have taken the RMSD same like the way done on VINA. We have analyze the comparison of VINA and PLANT and categorize the results like successful, good, better, and unsuccessful and I have named the protein on the basis of the results analysis.
3.2 DISCUSSION
The method has different significance over many others that predict protein–peptide interactions. One of the significances is that it does not require a known binding site. Such approach can specifically be tailored to predict the binding peptides and can be applied to any protein in case its structure is available. The ideas about binding peptides or proteins can be applied. Another significance is that it does not require that any substantial number of interactions be known for predictions to be made. The principle work of a single known or predicted peptide sequence can be applied. The method is supported through statistical statistic measure to estimate the reliability of predictions. The graphical representations reveal that the method can be applied to many structures systematically for identifying the strongest predictions, as well as making predictions as to whether binding occurs at all.
The approach applied has systemized what structural biologists often do when trying to guess a binding site from a protein surface. The experiment tried to match properties of a binding peptide with complementary properties on the protein surface. However, it has various significance such that these inferences are coupled to rationally derived knowledge of how amino acids, peptides, as well as proteins, can be determined by measure of the probability. Some predicted binding site can occur by chance. In case of such rare cases, it provides a more reliable starting point for site-directed mutagenesis or use of other studies designed for finding true binding sites.
The method applied is an excellent starting point for protein docking approaches. It revealed a fare and better reference method when applied to restricted binding regions instead of the entire protein surface. It is true that several sites are also found by a surface conservation method. However suc h observation is not surprising, since proteins that bind peptides will undoubtedly often show conservation of the peptide binding site. The inference is also true for all sites of molecular recognition. Tthe improved performance revealed in graphs show that such approaches indicates that the method offers a more precise, as well as specific ways of studying peptide binding sites as distinct from general functional sites. Nevertheless, the fact that when attempting to directly combining the two approaches the improvement in accuracy is marginal and the summed up averrage is high, revealing that there is still a complementary. The experiment shows that a predicted binding site is also conserved this can provide additional evidence for increasing confidence in a prediction.
Chapter 4: Conclusion
There is an increasing trend of developing peptide-based drugs using protein-peptide docking. The availability and accessibility of database of novel peptides determines the application of protein-peptide docking in development of peptide-based drugs. Despite the increasing trend of developing peptide-based drugs, the database of novel peptides is lacking. In this view, study has construct protein-peptide database with a view of using it in docking peptides into the target proteins. The main aim of the study is to create protein-peptide database of novel proteins involved in various therapeutic pathways. The creation of the protein-peptide database allows storage of novel peptides that can be retrieved and used in the design of peptide-based drugs. Furthermore, this study help to understand the application of protein-peptide docking in the design of peptide-based drugs. In this study docking of peptides into the target proteins has been done using VINA-PLANTS, a docking program and then further analyze results by comparing docking predictions with x-ray structures. The analysis of the study carried out by using Vega ZZ to calculate the RMSD results that that the program (VINA) was found successful to find the good peptide pose that binds the protein. And to further find out the best scored RMSD between VINA and Plant indicates that 74% (- negative) results indicating Vina is better than plant. In this regard the researchers should be smart enough to utilize this program in their overall comparison between Vina and Plant for their analysis of the good peptide.
In the post-genomic phase of biomedical research, the proteins, especially as target structures for many medicines, increasingly in the focal point of the scientific interest. To this extent, high-quality and quickly available Information on evolution, stability, dynamics and interactions of the proteins of eminently important. In this context, appropriate procedures and procedures were developed databases, which are used internationally. However the 3D structures of both the proteins as target structures, as well as the ligands play as lead structures for drugs. For the generation of reliable targeted Molecule structures have been developed.
The consideration of the Ligand flexibility via explicit conformers allowed the implementation to be faster procedures. This allowed for the in silico screening searchable 3D databases the millions of available natural products as well as synthetic substances contain (Preissner et al, 2011). One of us generated web-based drug database that has a fast similarity search the direct link to the medical target structures, through the inclusion of the WHO-defined indicator coding (ATC-codes) is widely used. The structure-assisted design of peptide libraries, also called nonlinear binding sites has been established. This results in new possibilities for the formation of peptidic binders proteins.
To determine the unfavorable pharmacological properties of the found peptides have become automated methods for the design of peptide mimetics developed. The use of conformational photo-switches both in biologically Active peptides as well as in small, non-peptide inhibitors is by us to develop procedures and opens up various new medical devices Applications. In addition to the development of bioinformatics methods, their application has been based on concrete, experimentally validated projects have contributed significantly to the success. It was Protein-protein interactions by peptide libraries are elucidated and ligands for different target molecules are proposed which are already in several cases patent applications.
As the number of known protein to protein interactions grows, so do the number of instances for which a peptide stretch is discovered to mediate interaction of importance. Similarly there is simultaneous increase of pace of structure determination of single proteins or domains. The outcomes shows that is rare to find globular domains lacking structural information. this therefore concludes that, taken together, this techniques is likely to describe the growing importance of peptides. Those interested in understanding, targeting as well as modifying protein interaction networks are supposed to critically analyze more biological processes.
REFERENCES
1. Bhachoo, J., &Beuming, T. (2017). Investigating Protein–Peptide Interactions Using the Schrödinger Computational Suite. Modeling Peptide-Protein Interactions: Methods and Protocols, 235-254.
2. Ciemny, M. P., Kurcinski, M., Kozak, K. J., Kolinski, A., &Kmiecik, S. (2017). Highly Flexible Protein-Peptide Docking Using CABS-Dock. Modeling Peptide-Protein Interactions: Methods and Protocols, 69-94.
3. Geng, C., Narasimhan, S., Rodrigues, J. P., &Bonvin, A. M. (2017). Information-Driven, Ensemble Flexible Peptide Docking Using HADDOCK. Modeling Peptide-Protein Interactions: Methods and Protocols, 109-138.
4. Goede, A., Dark, M., Mester, N., Frömmel, C., Preissner, R. Superdrug: aconformational drug database Bioinformatics. 21, 1751-3 (2005)
5. Goede, A., Preissner, R., Frömmel, C. Voronoi cell – a new error avoiding method For volume and density in proteins. J. Comp. Chem., 19, 1113-1123 (2010)
6. Gunther, S., Senger, C., Michalsky, E., Goede, A., Preissner, R. Representation oftarget-bound drugs by computed conformers: implications for conformational libraries. BMC Bioinformatics. 7, 293 (2006)
7. Kmiecik, S. (2017). Sebastian Kmiecik. Methods in Molecular Biology, 1484, 83-113.
8. Lee, H., &Seok, C. (2017). Template-Based Prediction of Protein-Peptide Interactions by Using GalaxyPepDock. Modeling Peptide-Protein Interactions: Methods and Protocols, 37-47.
9. Michalsky, E., Goede, A. and Preissner, “R. Loops” In Proteins (LIP) Comprehensive loop database for homology modeling. Protein Engineering 16, 1-7 (2003)
10. Preissner R. , Goede, A., Michalsky, E, Frömmel, C. Inverse sequence similarity in Proteins and its relation to the three-dimensional fold. FEBS Lett. 414, 425-429 (2011)
11. Preissner, R., Goede, A. and Frömmel, C. Dictionary of interfaces in proteins (DIP). Data bank of complementary molecular surface patches. J. Mol. Biol. 280, 535-550 (2012)
12. Zaidman, D., &Wolfson, H. J. (2017). Protein-Peptide Interaction Design: PepCrawler and PinaColada. Modeling Peptide-Protein Interactions: Methods and Protocols, 279-290.
Sherry Roberts is the author of this paper. A senior editor at MeldaResearch.Com in write my essay online if you need a similar paper you can place your order from write my essay for me services.
