With the module "Profiling", the user can retrieve information based on the identity of the chemical or its structure.
The Toolbox currently contains a large list of profiling methods (profilers) that identify the affiliation to previously defined categories, taking or not taking into account observed or predicted metabolites.
For most of the profilers detailed explanation is available.
AI. Profiling methods
Profiling methods (profilers) identify the affiliation to previously defined categories (e.g. the categories defined by the US-EPA for the assessment of new chemicals or the categories assessed within the OECD HPV Chemicals Program), mechanisms or modes of action. The outcome of the profiling also determines the most appropriate way to search for analogues.
Indeed not all profilers are relevant for all regulatory endpoints. While some general mechanistic profilers such as Protein binding are relevant for many endpoints (skin sensitization, acute toxicity, aquatic toxicity), some profilers are relevant for building categories for very specific endpoints only.
Profiling methods are separated into five groups based on the type of the qualitative information used for chemical assessment:
o Predefined
o General mechanistic
o Endpoint specific
o Empiric
o Toxicological
o Custom
The user can also create their own custom profiles. More details about creating custom profiles are available in section Creating a custom profiler.
1. Predefined
Predefined categories are set of rules for grouping chemicals developed by recognized institutions or organizations and they could be considered as “standard” categorization schemes.
The following predefined categorization schemes are implemented in the Toolbox:
o Database affiliation
o Inventory affiliation
o OECD HPV Chemicals categories
o Substance type
o US-EPA New Chemicals categories
2. General mechanistic
Mechanistic categorization schemes consist of rules for identification of important chemical characteristics based on published knowledge. Practically they rely on certain hypothesis about the studied phenomenon. In this respect their application is oriented for the phenomena for which they were developed.
The following general mechanistic categorization schemes are implemented in the Toolbox:
o DNA binding by OASIS
o DNA binding by OECD
o Estrogen receptor binding
o Protein binding by OASIS
o Protein binding by OECD
o Protein binding potency
o Superfragments
o Biodegradation probability (Biowin 1)
o Biodegradation probability (Biowin 2)
o Biodegradation probability (Biowin 5)
o Biodegradation probability (Biowin 6)
o Biodegradation probability (Biowin 7)
o Primary Biodeg (Biowin 3)
o Ultimate Biodeg (Biowin 4)
o Ultimate Biodeg
o BioHC Half-Life (Biowin)
o Hydrolysis half-life (Ka, pH7)
o Hydrolysis half-life (Ka, pH8)
o Hydrolysis half-life (Kb, pH7)
o Hydrolysis half-life (Kb, pH8)
o Hydrolysis half-life (pH=6.5-7.4)
o Ionization at pH=1
o Ionization at pH=4
o Ionization at pH=7.4
o Ionization at pH=9
o Toxic hazard classification by Cramer (original)
o Toxic hazard classification by Cramer (with extension)
3. Endpoint specific
The following endpoint specific categorization schemes are implemented in the Toolbox:
o Acute aquatic toxicity classification by Verhaar
o Acute aquatic toxicity MOA by OASIS
o Aquatic toxicity classification by ECOSAR
o Bioaccumulation – metabolism alerts
o Bioaccumulation – metabolism half-lives
o Biodegradation fragments (BioWIN MITI)
oDNA alerts for AMES, MN and CA by OASIS v.1.1
oDPRA Cysteine peptide depletion
oDPRA Lysine peptide depletion
oKeratinocyte gene expression
o Eye irritation/corrosion Exclusion rules by BfR
o Eye irritation/corrosion Inclusion rules by BfR
o Carcinogenicity (genotox and nongenotox) alerts by ISS
o in vitro mutagenicity (Ames test) alerts by ISS
o in vivo mutagenicity (Micronucleus) alerts by ISS
o Oncologic Primary Classification
o Skin irritation/corrosion Exclusion rules by BfR
o Skin irritation/corrosion Inclusion rules by BfR
4. Empiric
Empiric categorization schemes include rules used to determine the chemical elements constituting the target chemical, its chemical functionality and structural similarity.
The following empiric categorization schemes are implemented in the Toolbox:
o Chemical elements,
o Groups of elements,
o Lipinski Rule Oasis,
o Organic functional groups,
o Organic functional groups (nested),
o Organic functional groups (US-EPA),
o Organic functional groups, Norbert Haider (checkmol)
o Tautomers unstable
5. Toxicological
A repeated dose profiler used to identify the toxicological profiler of chemicals. The profiler contains boundaries based on repeated dose toxicity test data extracted from database of Hazard Evaluation Support System (HESS). The profiler is developed by NITE, METI (Japan) in cooperation with LMC
o Repeated dose (HESS)
6. Custom
Practically, these are the profiling schemes created by the user. In any case, while creating a new profiler the user may choose to place it in any of the above sections.
BII. Metabolism
The Toolbox is able to provide two types of metabolites, generated via:
All metabolites, documented and simulated, are searchable and profiled by the same approach as target chemical is thus informing the user about the potential formation of toxic metabolites of the target chemicals. Additional information concerning the studied species and experimental conditions is also provided by the databases.
1. Documented metabolism
Three databases of collected 400 observed metabolic pathways in microorganisms and mammals are implemented in the Toolbox.
· Observed microbial catabolism
Degradation pathways used by microorganism to obtain carbon and energy from 200 chemicals are stored in a special file format that allows easy computer access to catabolic information. The collection includes the catabolism of C1-compounds, aliphatic hydrocarbons, alicyclic rings, furans, halogenated hydrocarbons, aromatic hydrocarbons and haloaromatics, amines, sulfonates, nitrates, nitro-derivatives, nitriles, and compounds containing more than one functional group. Most of pathways are related to aerobic conditions. Different sources including monographs, scientific articles and public web sites such as the UM-BBD (L.B.M. Ellis, D. Roe, L.P. Wackett. Nucleic Acids Res., 34, D517 (2006); http://umbbd.msi.umn.edu/) were used to compile the database.
· Observed mammalian metabolism
Metabolic pathways documented for 100 chemicals with 630 studies in different mammals are stored in a database format that allows easy computer access to metabolic information. The collection includes chemicals with variety of functionalities, aliphatic amines, alkyl and aryl halides, ethers, esters, carbamates, carboxylic acid esters and multifunctional compounds. In vivo and in vitro studies were used to analyze the metabolic fate of chemicals. Metabolic maps with the in vivo studies are predominantly for the collection of studies (347 studies included in 49 maps). Microsomes prevail over the other experimental systems included in the in vitro studies (around 50% of studies). Around 50% of administration routes included the in vivo studies refer to oral route of administration. Different sources including monographs, scientific articles and public web sites were used to compile the database.
· Observed liver metabolism
Metabolic pathways documented for 447 chemicals in different mammals are stored in a database format that allows easy computer access to metabolic information. The collection includes chemicals with variety of functionalities, aliphatic hydrocarbons, alicyclic rings, furans, halogenated hydrocarbons, aromatic hydrocarbons and haloaromatics, amines, nitro-derivatives, and multifunctional compounds. In vivo and in vitro (predominantly microsomes) studies were used to analyze the metabolic fate of chemicals. Different sources including monographs, scientific articles and public web sites were used to compile the database.
2. Simulated metabolism
Simulators of molecular transformations imitating microbial, liver and skin metabolism, as well as abiotic hydroxylation are also implemented in the system. Below a short explanation for each of the simulators is presented:
· Autoxidation simulator
Autoxidation (AU) is a spontaneous, air-induced oxidation of organic molecules. It is a free- radical chain reaction of a chemical with molecular oxygen, resulting in the formation of oxidation products. Among the latter, organic hydroperoxides are regarded as the most important with respect to eliciting adverse effects such as contact allergy. An AU model was therefore developed to simulate the observed AU pathways. To this purpose, a training set of 133 chemicals (terpenes, simple aliphatic and polyethyleneglycol ethers, aldehydes, aminophenols) with published data relative to AU pathways. Data consistency was maintained by collecting data within the following test conditions: air or oxygen exposure, room temperature, atmospheric pressure, AU in bulk or in the presence of different solvents, nearly neutral (pH 7 - 7,5) or slightly alkaline (pH 8 - 9), medium, duration of AU from a few hours to several months.
· Hydrolysis (Acidic) simulator
Hydrolysis (Acidic) simulator predicts hydrolysis products of discrete organic chemicals under the following experimental conditions: acidic pH, ambient or room temperature and atmospheric pressure. The simulator was developed based on data collected from various sources, including articles and public web sites. The following classes of chemicals are included in the model : epoxides, aziridines, esters, carbamates, halomethanes, selected alkyl halides, anhydrides, dithiocarbamates, isocyanates, isothiocyanates, sulfonyl chloride, lactones, nitriles, amides, N-halamines, carbamates, diketenes, organic peroxides, etc.
· Hydrolysis (Basic) simulator
Hydrolysis (Basic) simulator predicts hydrolysis products of discrete organic chemicals under the following experimental conditions: basic pH, ambient or room temperature and atmospheric pressure. The simulator was developed based on data collected from various sources, including articles and public web sites. The following classes of chemicals are included in the model : sulfonyl halides, organophosphorus compounds, epoxides, aziridines, esters, carbamates, halomethanes, selected alkyl halides, anhydrides, dithiocarbamates, isocyanates, isothiocyanates, sulfonyl chloride, lactones, nitriles, amides, N-halamines, carbamates, diketenes, organic peroxides, etc.
· Hydrolysis (Neutral) simulator
Hydrolysis simulator is an abiotic model and predicts the hydrolysis products of chemicals and their quantities at 28 day under neutral or nearly neutral pH. The model predicts hydrolysis products of discrete organic chemicals under the following experimental conditions: neutral or nearly neutral pH, ambient or room temperature and atmospheric pressure. The simulator was developed based on data collected from various sources, including articles and public web sites.The following classes of chemicals are included in the model : discrete organic chemicals, epoxides, aziridines, esters, carbamates, halomethanes, selected alkyl halides, anhydrides, dithiocarbamates, isocyanates, isothiocyanates, sulfonyl chloride, lactones, nitriles, amides, N-halamines, carbamates, diketenes, organic peroxides, etc.
· Liver metabolism simulator
Multipathway modeling approach was used to simulate the metabolism in mammalian liver (O. Mekenyan, S. Dimitrov, R. Serafimova, E. Thompson, S. Kotov, N. Dimitrova, J. D. Walker 2004. Identification of the structural requirements for mutagencitiy, by incorporating molecular flexibility and metabolic activation of chemicals I. TA100 Model, Chemical Research in Toxicology , 17, 753 – 766). The scheme was conditioned by the fact that chemicals could be subject of variety of enzyme controlled reactions. Currently, 345 principal transformations are used to model metabolism in liver. They were separated into two major classes: non-rate determining and rate determining reactions. Transformations of highly reactive groups and intermediates such as acyl halide dehalogenation, geminal thiol halide dehalogenation, geminal halohydrine dehalogenation, N-nitrosoamine oxidative Ndealkylation, imide hydrolysis, sulfinic acid S-oxidation, etc. form the first class of molecular transformations. Various chemical equilibrium processes such as tautomerism are also included here. The second class includes oxidative, redox, reductive, hydrolytic and synthetic reactions.
Initially, the target chemical is submitted to the list of hierarchically ordered transformations. All transformations meeting the associated substructures are implemented on the parent producing the first level of metabolites. Each of the obtained metabolites is further submitted to the same list of transformations, thus producing the second level of metabolites, etc.
· Microbial catabolism simulator
The original CATABOL simulator of microbial metabolism is implemented in the system (J. Jaworska, S. Dimitrov, N. Nikolova, O. Mekenyan. SAR QSAR Environ. Res., 13, 307 (2002); S. Dimitrov, R Breton, D. Mackdonald, J. Walker, O. Mekenyan. SAR QSAR Environ. Res., 13, 445 (2002); S. Dimitrov, V. Kamenska, J.D. Walker, W. Windle, R. Purdy, M. Lewis, O. Mekenyan. SAR QSAR Environ. Res., 15, 69 (2004).). Single pathway catabolism is simulated using the abiotic and enzyme-mediated reactions via the hierarchically ordered principal molecular transformations extracted from documented metabolic pathway database. The hierarchy of the transformations is used to control the propagation of the catabolic maps of the chemicals. The simulation starts with the search for match between the parent molecule and the source fragment associated with the transformation having the highest hierarchy. If the match is not found search is performed with the next transformation, etc. When the match is identified, the transformation products are generated. The procedure is repeated for the newly-formed products. Predictability (probability that the metabolite is observed, given that the metabolite is predicted) evaluated on the bases of documented catabolism for 200 chemicals stored in the database of "Observed microbial catabolism" is 83%.
· Skin metabolism simulator
Skin metabolism simulator mimics the metabolism of chemicals in the skin compartment. Given the lack of reported skin metabolism data and the widespread hypotheses is that skin enzymes can metabolize absorbed xenobiotics via reactions analogous to those determined in liver, the simulator was developed as a simplified mammalian liver metabolism simulator. The skin metabolism simulator contains a list of hierarchically ordered principal transformations, which can be divided into two main types – rate-determining and non-rate-determining. The rate-determining transformations are Phase I and Phase II, such as C-hydroxylation, ester hydrolysis, oxidation, glutathione conjugation, glucuronidation, sulfonation. The non-rate-determining transformations include molecular transformations of highly reactive intermediates. The simulator starts by matching the parent molecule with the reaction fragments associated with the transformation having the highest probability of occurrence. This produces a set of first level metabolites. Each of these derived metabolites is then submitted to the same list of hierarchically ordered transformations, to produce a second level of metabolites. The procedure is repeated until a constraint for metabolism propagation is satisfied (e.g. low probability of obtaining a metabolite or application of Phase II reaction)
CIII. Explaining Profiling methods
There are three types profiling methods based on their structure:
· Standard profiling scheme
· Hierarchical type scheme
· Dendroidal type scheme
1. Standard type profiling scheme
To see an explanation of the categories from a profiling scheme, the user needs to select the scheme (1) and then to click on the button View (2). This will run the Profiling Scheme Browser (3). Initially, the Profiling Scheme Browser starts in Basic Mode. In this mode it will display references for each category definition from the profiler. Clicking on the button Advanced (4) will switch to Advanced Mode. (Figure 1)
Figure 1
In advanced mode the editor shows the query tree (1) and the content of the selected boundary and lets the user manipulate them. (Figure 2) New categories can be defined, existing ones can be renamed or deleted, rules can be added or deleted, new boundaries can be set. Changes for the profilers that are delivered via the Toolbox installation cannot be saved.
Figure 2
2. Hierarchical type profiling scheme
A new hierarchical organization of categories in profiling scheme is developed in Toolbox 3.0 (Figure 3). This organization is implemented for the following profiling schemes:
· DNA binding by OASIS
· DNA binding by OECD
· DNA alerts for AMES, MN and CA by OASIS v.1.1
· Protein binding by OASIS
· Protein binding by OECD
· Protein binding potency
· DPRA Cysteine peptide depletion
· DPRA Lysine peptide depletion
· Keratinocyte gene expression
Figure 3
3. Dendroidal type profiling scheme
The visualization of dendroidal type scheme is illustrated on Figure 4.
Figure 4
1 – logic panel: here a logic of the profiling scheme organized as a of logical nodes is presented
2- a panel with path report is given in panel 2.
3 – a panel with properties of nodes
4 – a panel with example structures that meets (does not meets) the criteria of the logical nodes is presented here.
AIV. Explaining Profiling results
Different organization of profiling results displayed on data matrix is available based on the three types profiling methods:
· Standard explanation
· Hierarchical explanation
1. Standard explanation
Profiling results obtained from standard type profiling schemes is shown on Figure 5:
Figure 5
Detailed explanation of the profiling result could be visualized when right click over the cell with profiling result (1) and selects Explain (2). (Figure 6)
Figure 6
Then the user has to select one of the available profiling results (1) and to click Details button (2). (Figure 7)
Figure 7
As a result of this action a window displaying the boundaries and mechanistic justification of the profiling method appears (1). (Figure 8)
Figure 8
Fields of the profiling results window are explained below (Figure 9):
Figure 9
1 – Panel with target chemical, if the target meets the criteria of the displayed category, then the target is highlighted in green.
2 – Structural boundary of the displayed category. If the target meets the criteria of the boundaries then it is marked with green tick.
3 – Structural fragment used to define the structural boundary of the displayed category
4 – A panel with Common fragments with elements included in the fragment used to define the diversity of substituents in the Structural fragment
5 – Mechanistic justification associated to each of the displayed category
There is a training set chemicals associated to the displayed categories. (6) (Figure 9). Training set chemicals are listed in panel training set. Usually structural boundaries are extracted based on these training set chemicals associated to categories. (Figure 10)
Figure 10
12. Hierarchical explanation
Based on this hierarchical type organization, the profiling results from obtained from these types scheme is illustrated on Figure 11:
Figure 11
Right click over cell (1) and select explain (2) to see details (4) for the selected category (3). (Figure 12)
Figure 12
A V. Creating a custom profile
In addition to the pre-installed categorization schemes, the user can create new profilers, implementing their own logic.
To do this, the user must click on the button New Scheme (1). This will bring up a dialog, which requires entering a name for the new categorization scheme. After entering the name, the system will create a new empty scheme and will load it in the Profiling Scheme Browser (2). (Figure 13)
Figure 13
1. Adding a new category
To add a new category the user must click on the button Add Category (1). After entering its name (2), the new category is created. Next, the user has to add its boundaries. (Figure 14)
Figure 14
2. Adding boundaries to a category
To add boundaries, the user must click on the button NEW (1). A list will appear (2) representing the five types of boundaries that describe a category:
· structural,
· parametric,
· referential,
· similarity and
· exact match.
Depending on the type of category selected, the user will be provided with different options to set the boundaries. For example for parametric boundaries limit values (minimum, maximum or both) are required, for structural boundaries structural fragment(s) are required. (Figure 15)
Figure 15
3. Combining category boundaries
In order to combine logically two or more boundaries, the user must create a query that describes the logic of the category scheme. Each set of boundaries is presented by a query clause (1). Query clauses could be inverted by using the button NOT (2). Additionally, the original clauses or the clauses that are derived from them, could be combined using logical buttons AND (3) or OR (4).
The picture below is presents a category containing two boundary sets - one parametric and one structural. Additionally these boundaries are combined in a way, which defines the following criteria:
- to have MW greater than 100 Da, AND
- NOT to have the fragment C(C)=O in their structure. (Figure 16)
Figure 16
A. VI. Export Profiling scheme
Toolbox allows the user to export profiling method. Select the desired scheme (1), open the profiling method (2), go to Advance mode (3), and then select Export scheme button (4). (Figure 17)
Figure 17
After this operation a window named Report Preview Form appears (1). The user could save the scheme in one of the following formats: PDF, HTML and RTF. The user has to click one of the displayed formats (2). (Figure 18)
Figure 19
AVII. Deleting a profiling scheme
The user is allowed to delete a custom profiler ONLY. The available profiling scheme, observed databases and simulators are not available for deletion.
To delete custom profiling scheme, please select the scheme that will be deleted (1), then click Delete scheme button (2). (Figure 20)
Figure 20