Introduction

Dengue is an important mosquito-borne disease caused by the dengue viruses (DENVs), which is estimated to cause around 25,000 deaths per year (Reginald et al. 2018). Aedes albopictus and Aedes aegypti are the commonly known vectors of transmission for DENV (Guardia and Lleonart 2014). At present, dengue is an endemic disease in more than 100 countries, especially in the sub-tropical and tropical regions (Chew et al. 2017). DENV-1-4 are the four serotypes of dengue virus, that are found to be antigenically as well as genetically distinct from each other. Surprisingly, even though they are distinct, they all cause similar sicknesses. People infected with any of these types may experience a range of symptoms, from asymptomatic fever to things like joint pain, rash, and other mild issues. It may also result in life-threatening symptoms like dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS) in severe cases (Shukla et al. 2023). The most important thing to note regarding these serotypes is that the infection with one induces lifelong immunity against only itself and not the other three serotypes. This is why the disease becomes of utmost concern. A common strategy that has been undertaken to stop DENV infection is primarily by vector control (Kala et al., 2023). The use of fog and the spreading of genetically modified (GM) mosquitoes are among the commonly known strategies that have failed to prevent the mosquito population from increasing. While active research on the development of vaccines has been ongoing for the past few decades, it has been held back by numerous challenges. The major problems faced during dengue vaccine development include the lack of appropriate animal models and the difference in antigenicity of each serotype which creates a hurdle in developing a single candidate against all four serotypes (Kala et al., 2023). The first dengue vaccine, CYD-TDV (chimeric yellow fever virus-tetravalent dengue vaccine) was a live-attenuated tetravalent vaccine comprising envelope proteins of DENV (Chew et al. 2017). However, its overall efficacy against DENV is low, especially that against DENV-2 remaining at only 39%. The CYD-TDV vaccine has also been shown to cause risks of hospitalization in children who are less than nine years of age. Thus, the World Health Organization (WHO) recommends the usage of this vaccine only in countries with an extremely high burden of dengue.

Also, there are no antiviral drugs against DENV available at present (Chew et al. 2017). The only therapy is supportive treatment with fluid restoration and close clinical monitoring. Although nucleoside analogs, like balapiravir, had entered preclinical and clinical trials, were terminated due to the lack of potency (Lim et al. 2013). Similarly, other anti-DENV drugs, such as chloroquine, celgosivir, and lovastatin, have undergone clinical trials but have failed to meet expectations (Lee et al. 2023). Unlike small molecules, peptides are marked by high selectivity and specificity, along with less off-target toxicity, which makes them insightful candidates for further testing. Due to their excellent pharmacological properties, several peptides are being tested against DENV (Lee et al. 2023). Also, the success of the antiviral peptide, Enfuvirtide, which has been recently approved by the Food and Drug Administration (FDA) for HIV/AIDS treatment, has further led the attention to antiviral peptides as an alternative to small molecule-based therapy or other treatment strategies for viral diseases (Diao and Meibohm 2013; Lee et al. 2023).

To combat DENV, various drug targets have been explored, focusing on both structural and non-structural proteins essential for viral maturation and replication. Structural proteins such as the envelope (E), capsid (C), and membrane (M) proteins offer potential targets due to their roles in viral assembly, membrane fusion, and RNA genome packaging (dos Santos Nascimento et al. 2022). Non-structural proteins like NS1, NS4A, NS4B, NS5, and the NS2B-NS3 complex are vital targets for drug design, given their involvement in viral replication, RNA synthesis, and modulation of host immune responses. Specifically, the NS2B-NS3 complex, with its catalytic triad essential for polyprotein cleavage, presents a promising target due to its vital role in viral replication and its conservation across DENV serotypes, making it an attractive focus for developing new therapeutic interventions (dos Santos Nascimento et al. 2022).

It is crucial to note that, nowadays the existence of data from certain resources helps to draw insights for various biological research and anti-viral research is one of them (Khan et al. 2021a, b, c, d). Without the data availability research may slow down. For the same reason, several specialized databases have been made in the past few years. Some of the most recent examples are ANTISTAPHYBASE (Zouhir et al. 2017), ANTIPSEUDOBASE (Zouhir et al. 2023), ACovPepDB (Zhang et al. 2022), HIPdb (Qureshi et al. 2013), CytomegaloVirusDb (Khan et al. 2021a, c, d), MMV-db (Khan et al. 2021a, c, d), HantavirusesDB (Khan et al. 2021a, c, d) which contain alternative therapeutic information (like peptide-based or RNA-based therapeutics) of traditional drugs against Staphylococcus species, Pseudomonas species, Coronavirus, HIV, Cytomegaloviruses, Mammarenaviruses, and Hantaviruses, respectively. To date, there is no single database specifically developed for anti-dengue virus peptides. The entire scenario of the spread of dengue worldwide and the limited treatment options, we were provoked to create a database specialized for anti-dengue peptides, named as Anti-Dengue Peptide Database (ADPDB). All the known antimicrobial peptides (AMPs) against DENV and related information have been curated in this crucial database. By providing a comprehensive resource, ADPDB aims to facilitate further research into anti-dengue peptides, potentially advancing therapeutic interventions against this challenging disease.

Drawing inspiration from recent advancements in anti-fungal and antiviral research, we recognize the importance of innovative strategies in combating infectious diseases. The application of subtractive proteomics and immunoinformatics in designing vaccine candidates against Candida auris, emphasizes the urgent need for effective therapeutic interventions (Khan et al. 2022a, b, c, d). Similarly, the significance of computational approaches in developing prophylactic vaccines against highly pathogenic coronaviruses reflects the broader landscape of infectious disease research (Khan et al. 2022a, b, c, d). These insights reinforce the importance of leveraging interdisciplinary methodologies and specialized databases, such as ADPDB, to accelerate the discovery and development of novel therapeutics targeting infectious agents like DENV.

Material & Method

Data Curation & Compilation

Curation of Anti-Dengue Peptides

PubMed literature database (https://pubmed.ncbi.nlm.nih.gov/) was used as a source database for curating the data. In PubMed, we used the following queries to curate the PubMed IDs of those articles which contain the information about anti-dengue peptides.

  • Query 1: ((((dengue virus) OR breakbone fever)) AND ((peptide) OR peptides)) AND ((inhibit*) OR block*).

  • Query 2: ((((dengue virus) OR dengue)) AND ((peptide) OR peptides)) AND ((inhibit*) OR block*).

The search of both the above queries was done on the 14th of October, 2023. Each query returned 819 PubMed IDs (total:1638). Further, all PubMed IDs were listed and the duplicate IDs were removed using Microsoft Excel. After getting rid of redundancy, we got 819 unique PubMed IDs. When the PMIDs of query 1 and query 2 and PMIDs after removing the duplicates were kept in an Excel file, we found that both queries returned the same PMIDs. Hence after removing redundancy, we got half of the total PMIDs of query 1 and query 2 (i.e., 819 out of 1638). All the PMIDs (returned by both queries and after removing duplication) are available in Table S1 of the supplementary file. After this, each article of respective PMID was thoroughly studied one by one and data was curated for the following fields: ‘ADPDB ID’, ‘anti-dengue peptide name’, ‘sequence’, ‘source’, ‘taxonomy’, ‘target organism with strain’, ‘inhibition concentration’, ‘target component’, ‘target process’, ‘mode of action (MoA)’, ‘toxicity’, ‘hemolytic activity’, and ‘validation’. Once a peptide was collected, we manually assigned an ID (e.g., ADPDB1) in ‘ADPDB ID’ field. It is important to note that in the case of some peptide sequences, we did not find a straightforward sequence (e.g., ‘FPFDFHHDRYYHFHWKRYQH’) instead we found the sequence with some special compounds (e.g., ‘Bz-Arg-Lys-L-Phg-NH2’) or the type of peptide (e.g., ‘cyclic peptide’) or a small description of peptide sequence (e.g., ‘Rhodanine-based peptide hybrid bearing a cyclohexyl moiety at the heterocycle’). These pieces of information were put as it is in the dataset. Moreover, in some cases ‘anti-dengue peptide name’ was missing and only the ‘sequence’ was available in the article and vice-versa. The missing values fields were filled up with a ‘not available’ string. After curating the data, we got our master anti-dengue peptide dataset with 606 peptides.

Compositional Details and Physicochemical Properties Computation

We utilized our custom Python package named proteinAnalysis2 (https://github.com/rajat-kumar-mondal/proteinAnalysis2) to calculate the compositional details and physicochemical properties of peptides. This package is developed in pure Python language (version 3.12.1) (https://www.python.org/), and an upgraded version of the original proteinAnalysis package (version 1) (Mondal et al. 2023). It incorporates functionalities from the proteinAnalysis class of the proteinAnalysis package (version 1) (Mondal et al. 2023), the ProteinAnalysis class of the ProtParam module of the Bio package (version 1.82) (https://biopython.org/), and the Peptide class of the peptides package (version 0.3.1) (https://pypi.org/project/peptides/). The proteinAnalysis2 package (https://github.com/rajat-kumar-mondal/proteinAnalysis2) is enormously capable of computing various basic details and properties of a peptide which are shown in Table 1.

Table 1 Table of compositional details and physicochemical properties that can be calculated by proteinAnalysis2 (https://github.com/rajat-kumar-mondal/proteinAnalysis2)
Full size table

By invoking the ‘all_comp_physP’ function from proteinAnalysis2 class of the proteinAnalysis2 package (https://github.com/rajat-kumar-mondal/proteinAnalysis2), users can effortlessly obtain a dictionary containing all the compositional details and physicochemical properties of a given peptide sequence. This function was specifically employed to compute these details for anti-dengue peptides. As we mentioned earlier in some cases of peptide sequence, we did not find the straightforward sequence and in some cases peptide sequence was unavailable, in those cases, we put the following information as a string.

  • If the peptide sequence is unavailable, we put “The peptide sequence is unavailable. Unable to calculate compositional details and physicochemical properties”.

  • If the peptide sequence contains some special characteristics (e.g., benzoyl compound), we put “Some special compound(s) is/are present with the peptide sequence. Unable to calculate compositional details and physicochemical properties”.

In this way, the computation of compositional details and physicochemical properties was done for 606 anti-dengue peptides that are present in our dataset.

Development of Backend and Frontend

The core of the ADPDB backend was constructed using an HTTPS server in PhpMyAdmin (version 5.1.1), housing a Relational Database Management System (RDBMS) i.e., MySQL server (version 5.7.36). The database schema of ADPDB is shown in Fig. 1. The frontend development employs HTML5, Bootstrap5, CSS3, JS, jQuery, Chart.js, DataTables, and AJAX.

Fig. 1
figure 1

Relational database schema of ADPDB. In the backend of the database there are 3 tables named ‘master_dataset’, ‘reference_details’, ‘compositional_physicochemical_details’. In the image the table names are shown in off-blue color background. All the attributes of the respective tables are shown in gray color background. All the tables are linked together by the ‘adpdb_id’ (underlined in red color) (primary key in the ‘master_dataset’ table), which act as a foreign key in table ‘reference_details’, and ‘compositional_physicochemical_details’

Full size image

User Interface of the Database

The user interface (UI) of the database is very simple, straightforward, responsive, and interactive. Figure 2 shows some of the glimpses of the database UI.

Fig. 2
figure 2

Some screenshots of ADPDB UI, where a. represents the homepage of ADPDB; b. represents the basic text search facility; c. represents the advanced search query builder; and d. represents look of an entry

Full size image

Homepage

The homepage of ADPDB displays a brief introduction about the database.

Search

On this page, a user will be able to find the simple text search and advanced search options. A user can search for any relevant information regarding anti-dengue peptides in the simple text search facility. In the case of the advanced search facility, a user can search in a specific field like ‘peptide name’, ‘length’, ‘source’, and more. A help button is given with both types of search facilities for user assistance.

Result Retrieving

All the search results will appear in a table format. The table is searchable and sortable itself. Pagination is also provided in the table for ease of use. Moreover, users can download specific entry/entries by selecting them or current page entries or all entries in a suitable format (including FASTA, TSV, CSV, LIST, JSON, TEXT, and Custom TSV) by clicking on the appropriate option from the download button. Users can view a single entry by clicking on the individual ID (e.g., ADPDB2) and export in FASTA, TEXT, and PDF format further.

Browse

On this page, the entire data of ADPDB is displayed. Users can study all the entries of ADPDB using this page.

Statistics

On this page, the data statistics are categorized by ‘peptide source’, ‘length’, ‘molecular weight’, and ‘net charge’ and displayed as an interactive bar chart. The Y axis of all bar charts is initially set to a scale of 10, as there is a considerable range between the lowest and highest number of records. This deliberate design choice aims to enhance visualizations. To adjust the scale, users can click on either the “Rescale X Axis” or “Rescale Y Axis” buttons to modify the plot accordingly. Click on “Rescale to Initial” to bring the plot to its initial point. Figure 3 shows the data statistics of ADPDB.

Fig. 3
figure 3

Data statistics of ADPDB. In the data statistics, a. represents the data distribution of all anti-dengue peptides based on the source, where most of the peptides (i.e., 428 peptides) source is synthetic, 1 from algae, 34 from animalia, 10 from bacteria, 4 from fungi, 16 from plante, 34 from virus, and there are 79 peptides for which any specific source were not found; b. represents the data distribution of all anti-dengue peptides by length, where most of the peptides (i.e., 81 peptides) fall between the length of 11 to 20 amino acid residues, 64 peptides between the length of 1 to 10 amino acid residues, 41 peptides between the length of 21 to 30 amino acid residues, 22 peptides between the length of 31 to 40 amino acid residues, 5 peptides between the length of 41 to 50 amino acid residues, 6 peptides between the length of 51 to 60 amino acid residues, and 1 peptide between the length of 71 to 80 amino acid residues; c. represents the data distribution of all anti-dengue peptides based on molecular weight, where most of the peptides (i.e., 67 peptides) fall between 1000 to 2000 Dalton (Da), 45 peptides between 1 to 1000 Da, 45 peptides between 2000 to 3000 Da, 42 peptides between 3000 to 4000 Da, 13 peptides between 4000 to 5000 Da, and 8 peptides between 5000 to 10,000 Da; d. represents the data distribution of all anti-dengue peptides based on their net charge, where the net charge of most of the peptides (i.e., 112 peptides) lies between 0 to 5, 2 peptides between -10 to -5, 73 peptides between -5 to 0, 27 peptides between 5 to 10, 4 peptides between 10 to 15, and 2 peptides between 15 to 20. In the case of representation of b., c., and d., there were 386 peptides for which length, molecular weight, and net charge could not be computed due to the unavailability of the sequence or availability of special characteristics

Full size image

Downloads

From this page, a user can download the master dataset of ADPDB in XLSX, CSV, TSV, LIST, TEXT, and FASTA format.

News/Updates

All the latest news/updates regarding ADPDB will be displayed on this page.

Developers

Users can find all the developer’s information on this page.

Help

On this page, all the details of the search facilities and a full description of the TEXT format of ADPDB are shown.

Contact

The contact details of the principal investigator and core technical developers are displayed here. Users can reach out to them from this page. A web form is also provided on the same page which can be used by the user to raise any query/doubt.

In Fig. 4, a diagrammatic representation of the end-to-end development of ADPDB is shown.

Fig. 4
figure 4

Diagrammatic representation of end-to-end development process of ADPDB. The process begins with identifying relevant literature in the PubMed database using keywords. The flowchart shows that 1638 PubMed IDs were extracted based on a keyword search. Next, 819 unique PubMed IDs containing anti-dengue peptides were identified. Manual data curation from these articles is then performed to create a master dataset of 606 anti-dengue peptides, followed by the computation of compositional details and physicochemical properties of the peptides. Then, an interactive GUI for the database was developed. Finally, ADPDB compared with other existing databases

Full size image

Results and Discussion

Insights from ADPDB

ADPDB is the first comprehensive knowledge base of anti-dengue peptides found widely. This database can be specifically used to target dengue viruses (including DENV-1, DENV-2, DENV-3, DENV-4). The database supports simple text searches to intuitive advanced search facility. Moreover, the database also allows user to manipulate the data as per their need and export it in multiple formats like TSV, TEXT, LIST, FASTA, JSON, etc. in local machines for further analysis or development. Users also can obtain fully customized reports as per their requirements from the database’s custom report facility. While viewing an individual entry the database displays the result in the following 6 parts.

  1. 1.

    General description which includes ADPDB ID, peptide name, source, taxonomy, and validation (i.e., whether the peptide is validated experimentally or in silico).

  2. 2.

    Peptide sequence & composition which includes length, molecular formula, amino acid (AA) counts, AA frequencies, missing AA, AA which occurs most & less, hydrophobic & hydrophilic AA counts, acidic & basic AA counts, counts and frequencies of modified AA.

  3. 3.

    Physicochemical properties which include molecular weight, aromaticity, aliphatic index, instability index, hydrophobic moment, GRAVY, net charge, secondary structure fraction (SSF), molar extinction coefficient, mass shift, etc.

  4. 4.

    Structural class of the peptide.

  5. 5.

    Activity information includes target organism, family, inhibition concentration information, target component information, target process information, mode of action, toxicity, and hemolytic activity.

  6. 6.

    Database cross-references which include the PubMed ID (this is the PubMed ID of that literature from where the data is collected for the respective ADPDB entry).

At present, this database contains a total of 606 peptide entries. These peptides originate from different sources: Algae (1), Animalia (34), Bacteria (10), Fungi (4), Plantae (16), Synthetic (428), Virus (34). There are 79 peptides for which the source is not available. Out of 606 entries, 32 peptide sequences are unavailable in the database due to their unavailability in the source literature, 352 sequences are not containing straightforward peptide sequence, and 222 sequences are pure peptide sequences (do not contain any special compound with them) that are present in the database. The length of the peptides varies between 1 and 80 AA, molecular weight lies between 1 and 10,000 Da, and net charge lies between -10 and 20. Scientists, researchers, and pharmaceutical industries can use this database for novel drugs and therapeutics development for dengue.

Recent studies have shown the utility of bioinformatics tools and computational approaches in vaccine design against human pathogens. These approaches enable the identification of potential vaccine targets, mapping of immune epitopes, and design of effective vaccine candidates (Khan et al. 2022a, c, d). Similarly, ADPDB utilizes bioinformatics strategies to curate and analyze anti-dengue peptides. Like the studies mentioned, ADPDB serves as a valuable resource for researchers and pharmaceutical industries, providing comprehensive information on anti-dengue peptides. Researchers can leverage ADPDB to identify potential peptide-based therapeutics, study structure-activity relationships, and accelerate the development of novel drugs for combating the dengue virus. Thus, ADPDB aligns with the broader landscape of infectious disease research, facilitating the discovery and development of therapeutics against the dengue virus through bioinformatics-driven approaches.

Comparison with Other Existing Databases

Till now, many databases have been dedicated to AMPs in a generalized and specialized manner, since antimicrobial resistance (AMR) is a major issue. Some of the well-known generalized databases include CAMP R4 (Gawde et al. 2023), DRAMP 3.0 (Shi et al. 2022), dbAMP 2.0 (Jhong et al. 2022), AMPDB v1 (Mondal et al. 2023), APD3 (Wang et al. 2016), etc. There are also renowned specialized databases of AMPs like Peptaibols database (Whitmore and Wallace 2004), Defensins knowledgebase (Seebah et al. 2007), LAMP (Zhao et al. 2013), InverPep (Gómez et al. 2017), MilkAMP (Théolier et al. 2014), PhytAMP (Hammami et al. 2009), RAPD (Li et al., 2008), SAPdb (Mathur et al. 2021), ANTISTAPHYBASE (Zouhir et al. 2017), ANTIPSEUDOBASE (Zouhir et al. 2023), DRAVP (Liu et al. 2023), AVPdb (Qureshi et al. 2014), HIPdb (Qureshi et al. 2013), ACovPepDB (Zhang et al. 2022).

The generalized AMP databases contain overall every type of AMP whereas the specialized AMP databases contain the data of some specific types. Peptaibols database (Whitmore and Wallace 2004) and Defensins knowledgebase (Seebah et al. 2007) contain the data of peptaibols and defensins respectively whereas LAMP (Zhao et al. 2013) deals with linking AMPs. InverPep (Gómez et al. 2017), MilkAMP (Théolier et al. 2014), and PhytAMP (Hammami et al. 2009) contain data of invertebrate AMPs, AMPs from milk sources, and plant AMPs respectively. RAPD (Li et al., 2008) and SAPdb (Mathur et al. 2021) deal with the data of recombinant and synthetic AMPs respectively. DRAVP (Liu et al. 2023) and AVPdb (Qureshi et al. 2014) have information about the antiviral peptides. ANTISTAPHYBASE (Zouhir et al. 2017) and ANTIPSEUDOBASE (Zouhir et al. 2023) are two specific target bacteria databases that contain data about AMPs and essential oils (EOs) that can be used to target Staphylococcus and Pseudomonas species, whereas HIPdb (Qureshi et al. 2013) and ACovPepDB (Zhang et al. 2022) are only two target virus databases that holds that information about anti-HIV and anti-Corona virus peptides specifically.

All the databases (generalized and specialized) that are dedicated to AMPs, to date, hold very useful information for antimicrobial drugs and therapeutics developments. However, there is no single database which is dedicated to anti-dengue peptides. Moreover, dengue is very prevalent in tropical and sub-tropical countries and the mortality rate for dengue is quite high (Chew et al. 2017). The drugs that are available in the market for dengue are less effective and often produce side effects (Obi et al. 2021). In this scenario, anti-viral peptides (a type of AMP) can be a better therapeutic alternative. Keeping all this in mind, we developed the ADPDB, the first dedicated database for anti-dengue peptides, to develop a better, safer, and economical treatment strategy for the dengue virus. At this moment, ADPDB does not contain any clinical trial information about anti-dengue peptides, sequence alignment tools (like BLAST for comparing a query sequence to ADPDB sequences to find regions of local similarity; or MUSCLE to align multiple sequences of ADPDB to identify regions of similarity and potential functional or evolutionary relationships) or anti-dengue peptide prediction tool or any other related tools. These are the few limitations of ADPDB. However, it will be improved in its future versions.

Conclusion & Future Perspectives

Despite an approved dengue vaccine, it doesn’t cover all DENV types, posing a challenge in affected regions. Urgently needed are effective drugs for all DENV strains. Promisingly, anti-dengue peptides offer a solution, serving as both drugs and vaccine candidates. While biocompatible and cost-effective, efforts to enhance their effectiveness, using long peptides and nanoparticle delivery, are underway. Recognizing the need, we’ve developed ADPDB, a dedicated anti-DENV peptide database.

ADPDB represents a significant advancement in the field of infectious disease research, particularly in combating the dengue virus. With dengue posing a substantial global health challenge and limited treatment options available, ADPDB emerges as a comprehensive knowledgebase dedicated to anti-dengue peptides and right now it holds 606 anti-dengue peptides. By consolidating information on peptides exhibiting anti-dengue activity sourced from extensive literature curation, ADPDB offers a valuable resource for researchers, pharmaceutical industries, and clinicians. ADPDB not only addresses the critical gap in specialized databases focusing on anti-dengue peptides but also aligns with the growing interest in peptide-based therapeutics. Its user-friendly interface, offering functionalities such as simple and advanced search options, data retrieval, customizable reports, and data statistics, enhances accessibility and usability. Using information from ADPDB, the researchers can design new drugs using studies like structure-activity relationships or generate new leads by incorporating the data into ML models.

In comparison to existing databases dedicated to AMPs, ADPDB stands out as the first of its kind focusing specifically on anti-dengue peptides. Given the prevalence of dengue in tropical and sub-tropical regions and the urgency for effective treatment strategies, ADPDB holds immense promise in advancing therapeutic interventions against the dengue virus. Despite its current limitations such as the absence of clinical trial information and certain analytical tools, ADPDB sets a solid foundation for future improvements and enhancements. As efforts continue to evolve, ADPDB is poised to foster collaboration, innovation, and the identification of novel therapeutic strategies in the global fight against dengue.

All in all, ADPDB represents a pivotal resource that not only consolidates existing knowledge but also propels the field forward, offering hope for improved treatments and ultimately contributing to the mitigation of the dengue virus’s significant impact on human health worldwide.