Home Database Statistics Download Help
SkinSensDB: a curated database for skin sensitization assays
Skin sensitization is an important toxicological endpoint in drug development and regulatory decision making. Chemical sensitizers act as haptens binding to protein molecules to trigger immune responses that could induce allergic contact dermatitis. To facilitate development of AOP-based computational prediction methods, a novel curated database named SkinSensDB has been constructed by manual curation of published literatures.
AOP and assays
Adverse Outcome Pathway and associated assays
More information in AOPwiki
News

[2017/10/19]
[UPDATE] data from PMID:22659254 have been updated and fixed
[2017/10/02]
[FUNCTION] New models supporting integrated testing strategy
[2017/08/10]
[UPDATE] 131 human data from Basketter et al., 2014
[2017/03/29]
Server is back online
[2017/03/23]
Server is down for upgrading

[Archived News]
Citation

Chun-Wei Tung*, Chia-Chi Wang, Shan-Shan Wang (2017) Mechanism-informed Read-across Assessment of Skin Sensitizers Based on SkinSensDB. (submitted)

Chia-Chi Wang, Ying-Chi Lin, Shan-Shan Wang, Chieh Shih, Yi-Hui Lin, Chun-Wei Tung* (2017) SkinSensDB: a curated database for skin sensitization assays. Journal of Cheminformatics. 9, 5.

Contact

Chun-Wei Tung, Ph.D.
Associate Professor
School of Pharmacy & PhD Program in Toxicology,
Kaohsiung Medical University
E-mail: cwtung@kmu.edu.tw
Webpage: http://cwtung.kmu.edu.tw

Export to Excel
Search
Draw structure Convert from SMILES
search
Exact Search Substructure Search Similarity Search
Predict
Draw structure Convert from SMILES
search
Predict
Browse

Please wait...

Search Clear File download
{{item.Name}} No matches found. {{item.CasNo}} No matches found. {{item.CID}} No matches found. {{tox.label}} {{tox.label}} {{tox.label}} {{tox.label}} {{tox.label}}
{{e.Name}} {{e.CasNo}} {{e.CID}} report verified_user help ({{toFixed(e.DPRApRatio,2)}}) report verified_user help ({{toFixed(e.KeratinoSenspRatio,2)}}) report verified_user help ({{toFixed(e.hclatpRatio,2)}}) report verified_user help insert_chart ({{toFixed(e.LLNApRatio,2)}}) report verified_user help insert_chart
{{searchWay}} - {{smile}}

Please wait...

No hits found!
Your search gets no results!
{{item.Name}} No matches found. Similarity Score {{item.CasNo}} No matches found. {{item.CID}} No matches found. {{tox.label}} {{tox.label}} {{tox.label}} {{tox.label}} {{tox.label}}
{{e.Name}} {{e.score | number}} {{e.CasNo}} {{e.CID}} report verified_user help ({{toFixed(e.DPRApRatio,2)}}) report verified_user help ({{toFixed(e.KeratinoSenspRatio,2)}}) report verified_user help ({{toFixed(e.hclatpRatio,2)}}) report verified_user help insert_chart ({{toFixed(e.LLNApRatio,2)}}) report verified_user help insert_chart
Results - {{Name|limitTo:20}} ...
Basic Info

Please wait...

NAME {{Name}}
CAS No. {{casno}}
No structure information available

Structure

Pubchem {{basic.CID}}

Properties

NAME {{Name}}
CAS No. {{casno}}
IUPAC NAME {{basic.IUPACName}}
IUPAC INCHI {{basic.InChI}}
IUPAC INCHIKEY {{basic.InChIKey}}
Molecular Formula {{basic.MolecularFormula}}
Canonical SMILES {{basic.CanonicalSMILES}}
Isomeric SMILES {{basic.IsomericSMILES}}
H-bond Acceptor {{basic.HBondAcceptorCount}}
H-bond Donor {{basic.HBondDonorCount}}
Molecular Weight {{basic.MolecularWeight}}
TPSA {{basic.TPSA}}
External trnk PubChem: CID{{basic.CID}}
2D structure 3D structure
DPRA/PPRA [ {{dList.length}} record(s) ]
report verified_user help {{d.PMID}} {{d.Assay}} {{d.HRP}} {{d.Cys_PeptideCon}} {{d.Cys_PeptideConUnit}} {{d.Cys_ChemicalCon}} {{d.Cys_ChemicalConUnit}} {{d.Cys_Result}} {{d.Cys_SD}} {{d.Lys_PeptideCon}} {{d.Lys_PeptideConUnit}} {{d.Lys_ChemicalCon}} {{d.Lys_ChemicalConUnit}} {{d.Lys_Result}} {{d.Lys_SD}} {{d.Glu_PeptideCon}} {{d.Glu_PeptideConUnit}} {{d.Glu_ChemicalCon}} {{d.Glu_ChemicalConUnit}} {{d.Glu_Result}} {{d.Glu_SD}} {{d.His_PeptideCon}} {{d.His_PeptideConUnit}} {{d.His_ChemicalCon}} {{d.His_ChemicalConUnit}} {{d.His_Result}} {{d.His_SD}} {{d.GSH_PeptideCon}} {{d.GSH_PeptideConUnit}} {{d.GSH_ChemicalCon}} {{d.GSH_ChemicalConUnit}} {{d.GSH_Result}} {{d.GSH_SD}}
KeratinoSens/LuSens [ {{kList.length}} record(s) ]
report verified_user help {{d.PMID}} {{d.Assay}} {{d.ImaxFoldinduction}} {{d.EC15}} {{d.conUnitEC15}} {{d.IC50}} {{d.conUnitIC50}} {{d.EC2}} {{d.conUnitEC2}} {{d.EC3}} {{d.conUnitEC3}}
h-CLAT [ {{hList.length}} record(s) ]
report verified_user help {{d.PMID}} {{d.CD86_EC150}} {{d.CD86_EC150_unit}} {{d.CD86_EC150_MAXRFI}} {{d.CD54_EC200}} {{d.CD54_EC200_unit}} {{d.CD54_EC200_MAXRFI}} {{d.CV75}} {{d.CD86_EC150_unit}}
LLNA [ {{LList.length}} record(s) ]
report verified_user help {{d.PMID}} {{d.ChemicalClass}} {{d.LLNAVehicle}} {{d.EC3}} {{d.Conc1}} {{d.SI1}} {{d.Conc2}} {{d.SI2}} {{d.Conc3}} {{d.SI3}} {{d.Conc4}} {{d.SI4}} {{d.Conc5}} {{d.SI5}} {{d.Conc6}} {{d.SI6}} {{d.Conc7}} {{d.SI7}} {{d.Conc8}} {{d.SI8}}
Human Data [ {{humanList.length}} record(s) ]
report verified_user help {{d.PMID}} {{d.HumanClass}} {{d.CD86_EC150_unit}} {{d.HumanNOEL}}
Content

SkinSensDB vertical_align_top Back to content

SkinSensDB is a web-based resource providing useful information of structures, physicochemical properties and experimental data from skin sensitization assays. Browse and search tools were implemented to facilitate the exploration of skin sensitization data. The integration of chemical structures, physicochemical properties and experimental results from AOP-related assays could be helpful for the development of a AOP-based prediction system integrating all models corresponding to the four major events.

History vertical_align_top Back to content

[2017/10/19] v1.5 [UPDATE] data from PMID:22659254 have been updated and fixed
[2017/10/02] v1.4 [FUNCTION] New models supporting integrated testing strategy
[2017/08/10] v1.3 [update] 131 human data from Basketter et al., 2014
[2017/01/31] Paper published
[2017/01/19] V1.2 Update & bug fix
[2016/12/30] V1.1 Update
[2016/03/04] V1.0 Online

Classification Criteria vertical_align_top Back to content To give an intuitive overview of the experiment results, the database automatically assigns chemicals as sensitizers/nonsensitizers based on the results from various assays using the following classification criteria derived from OECD test guidelines. Please note that replicate test results of h-CLAT required by the OECD test guideline are rarely reported in publications. Hence, in this database, the positives for h-CLAT are classified based on one test rather than two replicate tests.
DPRA/PPRA Reactivity <=6 .38% Negative
Reactivity > 6.38% Positive
KeratinoSens/Lusens EC1.5 >= 1000 uM Negative
EC1.5 < 1000 uM Positive
hClat Neither CD86 EC150 nor CD54 EC200 was determined Negative
CD86 EC150 <= CV75 or CD54 EC200 <= CV75 Positive
LLNA SI < 3 Negative
SI >= 3 Positive
Human data Category 5-6 Negative
Category 1-4 Positive
Search SkinSensDB vertical_align_top Back to content

Three kinds of search functions are also available in SkinSensDB including exact, substructure and similarity searchs. All the search functions were implemented using RDKit library. For the input of chemical structures, users can either draw the chemical structure of interest or enter the SMILES string for converting to a chemical structure. The user interface for drawing chemical structures and entering SMILES string is shown in Figure 5 and 6, respectively.


Figure 5. The user interface of search function by drawing a chemical structure. This function is based on a JSME editor.


Figure 6. The tool for inputing chemcial structures by converting from a SMILES string. Step 1: enter the SMILES string. Step 2: click the search icon. The page will be redirected to the JSME editor as shown in Figure 5 with chemical structure appeared.

The search result will be shown in the form of table the same as the Browse tool with full functions. To access the search results, a tab showing the search type and SMILES string will be available just adjacent to the Browse tool. Figure 7 and 8 are example results for substructure and similarity searchs, respectively. For exact and substructure search, only chemicals fitted to the selection rules will be shown in the table. For similarity search, all chemicals will be shown in the resulted table with an additional column showing the similarity between the query and target chemcials.


Figure 7. The example page showing the results of a substructure search.


Figure 8. The example page showing the results of a similarity search. There is an additional column representing the similarity between query and target chemicals. The table is sorted by the similarity for users to find the most similar chemicals.

Batch query vertical_align_top Back to content

To perform a batch query, users can use the search box in the Browse tool.


Figure 9. An example of a batch query showing the filtered results. Users can enter multiple CAS numbers and click on the button with an arrow icon to perform a batch query. The button with a trash can icon can be utilized to reset the filter. The button with a download icon can be utilized to export the summary information as an EXCEL file.

Submit data vertical_align_top Back to content

Contributions to SkinSensDB is highly welcomed. Users who wish to submit their data can use the following Excel file to submit data to cwtung@kmu.edu.tw. [Download Submit Form]

AOP vertical_align_top Back to content

AOP (Adverse outcome pathway) represent the molecular initiating event (MIE), key events and outcome associated with an adverse effect. For Skin sensitization, its molecular initiating event is the covalent binding of chemicals to protein following three key events of keratinocyte activation, activation of dendritic cells, and T-cell activation/proliferation. As shown in the following figure, assays associated with individual events are listed.
For more information, please refer to AOPwiki at https://aopwiki.org/wiki/index.php/Aop:40

Figure 10. The AOP of skin sensitization. A skin sensitizer will induce the four key events to cause sensitization. First, a chemical should covalently bind to protein marcromolecules for recognition by immune systems. Second, keratinocyte should be activated by the chemical-protein complex. Third, dedritic cells should be activated. Finally, T-cell proliferation should be induced. The adverse outcome of the four-step AOP pathway is allergic contact dermatitis.

Integrative testing strategy models vertical_align_top Back to content

Two intuitive and interpretable models including a majority vote (2-out-of-3) model (Urbisch et al., 2014) and a decision tree model (Roberts and Patlewicz, 2017) were developed in this study. For the majority vote model, a 2-out-of-3 rule will be applied to evaluate results from alternative assays of three AOP events of protein binding, keratinocyte activation and activation of dendritic cells to determine the potential of skin sensitization for test chemicals. In contrast, the decision tree model considers only two AOP events of protein binding and activation of dendritic cells. The decision tree model can be summarized as a simple rule of 'IF one of the two events is positive THEN the chemical is a sensitizer ELSE the chemical is a non-sensitizer'.

Read-across assessment of skin sensitization vertical_align_top Back to content

The read-across tool (Figure 11) can be accessed from both the 'Predict' tool (Figure 12) and the 'Browse' function (Figure 13). The read-across analysis will appear as an interactive panel showing the most similar chemicals with similarity scores and corresponding experimental data. Integrative testing strategy models will be applied to predict the potential of skin sensitization.

Figure 11. The interative panel showing the read-across prediction of skin sensitization potential.

Users can input a chemical by drawing the structure or converting from SMILES (Figure 12).

Figure 12. The interface of read-across prediction function.

From the Browse tool, a model icon will appear when there is no corresponding LLNA or human data for a specific chemical (Figure 12). By clicking the model icon, an interactive panel (Figure 13) will show the analysis results for the selected chemical. In contrast to the prediction function searching structurally similar chemicals for all three AOP events, the modified browsing tool combines assay results curated from literature and read-across analyses for events without curated experimental data.

Figure 13. The model icon can be accessed from the Browse tool.

Citation vertical_align_top Back to content

Chia-Chi Wang, Ying-Chi Lin, Shan-Shan Wang, Chieh Shih, Yi-Hui Lin, Chun-Wei Tung* (2017) SkinSensDB: a curated database for skin sensitization assays. Journal of Cheminformatics. 9, 5.

Reference vertical_align_top Back to content

  • [NICEATM] (2013) NICEATM LLNA database.
  • [RDKit] RDKit: Open-source cheminformatics.
  • [JSME] Bienfait B, Ertl P (2013) JSME: a free molecule editor in JavaScript. J Cheminformatics. doi: 10.1186/1758-2946-5-24
  • [PubChem-PUG] Kim S, Thiessen PA, Bolton EE, Bryant SH (2015) PUG-SOAP and PUG-REST: web services for programmatic access to chemical information in PubChem. Nucleic acids Res 43:W605-W611. doi: 10.1093/nar/gkv396
  • [Majority vote model] Urbisch et al. (2015) Assessing skin sensitization hazard in mice and men using non-animal test methods. Regul. Toxicol. Pharmacol. 71, 337-351, doi:10.1016/j.yrtph.2014.12.008.
  • [Decision tree model] Roberts DW, Patlewicz G (2017) Non-animal assessment of skin sensitization hazard: Is an integrated testing strategy needed, and if so what should be integrated? J. Appl. Toxicol. doi:10.1002/jat.3479..
  • Content

    Overview
    Number of unique compounds
    {{ chart1 }}
    Number of bioactivity data
    {{ chart2 }}
    vertical_align_top Back to content
    Distribution
    Distribution of peptide depletion of DPRA/PPRA
    {{ chart3 }}
    Distribution of bioactivity data of h-CLAT
    {{ chart5 }}
    Distribution of bioactivity data of KeratinoSens/LuSens
    Distribution of bioactivity data of LLNA
    Distribution of human data
    vertical_align_top Back to content
    Classification
    DPRA/PPRA
    {{ chart7 }}
    KeratinoSens/LuSens
    {{ chart8 }}
    h-CLAT
    {{ chart9 }}
    LLNA
    {{ chart10 }}
    Human Data
    {{ chartHumanSkinsens }}
    vertical_align_top Back to content
    Dataset for downloading
    In addition to the summarized data table (as shown in the [Browse] page) providing binary results (i.e. positive or negative) of assays for evaluating skin sensitizers, the assay values could also be utilized as a reference dataset for developing prediction models, such as the quantitative structure-activity relationship (QSAR) models. To facilitate the development of prediction models, the data table has been exported as a tsv (tab-delimited value) file and is downloadable via the following link.
    The file contains information of Chemical Name, CAS No., PubChem CID, Canonical SMILES, the highest value of DPRA/PPRA assays, the lowest EC15 value of KeratinoSens/LuSens assays, the lowest EC and CV75 values of h-CLAT assays, the EC3 value of LLNA assays and human data.
  • [2017-10-19] - data-20171019.tsvfiber_new
  • [2017-01-19] - data-20170119.tsv