| SEP | |
|---|---|
| Title | Representation of Parts and Devices for Build Planning |
| Authors | Jacob Beal (jakebeal@ieee.org), Vinoo Selvarajah, Gael Chambonnier, Traci Haddock-Angelli, Alejandro Vignoni, Gonzalo Vidal, Nicholas Roehner |
| Editor | |
| Type | Data Model |
| SBOL Version | SBOL 3.1 |
| Replaces | None |
| Status | Draft |
| Created | 6-Jan-2022 |
| Last modified | 17-Feb-2022 |
| Issue | #113 |
This SEP proposes a set of terminology and practices for representing genetic parts and functional devices at various stages of design, synthesis, and assembly. These practices are intended to represent any of the wide array of approaches based on embedding parts in carrier vectors, such as BioBricks, Gateway, MoClo, GoldenBraid, PhytoBricks, and other Type IIS methods.
- 1. Rationale
- 2. Specification
- 3. Example or Use Case
- 4. Backwards Compatibility
- 5. Discussion
- 6. Competing SEPs
- References
- Copyright
There are often multiple steps to creating and documenting a genetic product from simpler building blocks in a design. For example, a fragment of DNA may be synthesized as an insert into a vector backbone, then digested out of that backbone and assembled together with other fragments to produce a final construct. It is often unclear which stage a given sequence is associated with.
While SBOL has all of the tools necessary to represent such stages and their relationships, no set of recommended terminology and practices has yet been proposed to standardize around. This document aims to propose a set of best practices that are both succinct and applicable to produce unambiguous descriptions of DNA sequences at every stage of design, synthesis, and assembly for a wide variety of approaches based on embedding parts in carrier vectors, such as BioBricks, Gateway, MoClo, GoldenBraid, PhytoBricks, and other Type IIS methods.
The approaches recommended in this proposal may be applicable to other types of molecules besides DNA and methods of building, but those applications are beyond the scope of this proposal.
All of the activities on the following non-exhaustive list are intended to be supported cleanly with this proposal:
- Identify whether a sequence describes a construct in a vector, after extraction from a vector, after assembly with other constructs, etc.
- Predict the sequence of an assembled construct from the components to be assembled, including scars and drop-outs.
- Determine a plan for assembly of a composite constructs from simpler constructs.
- Produce the same plasmid by either synthesis or assembly
- Determine sequences for a synthesis order
- Determine whether a construct needs to have flanking sequences added to it to enable assembly
- Check whether a construct containes digestion sites that are incompatible with an assembly plan
- Maintain a catalog for a distribution of constructs in backbones.
- Maintain a catalog of the same construct stored in multiple different backbones
- Maintain a catalog of the same construct stored with multiple different flanking sequences for incompatible assembly methods.
This proposal does not propose any changes to the SBOL data model. Rather, it is intended to only be a set of best practices to follow.
This specification and set of practices for representation of parts and devices for build planning uses the following definitions:
-
Part: Design for a single contiguous linear DNA construct with a completely specified sequence.
-
Unitary Part: Any part that is not designed with reference to an assembly. In many cases a unitary part may also be a device with a function that can be defined simply (e.g., promoter, CDS, terminator), but unitary parts can potentially also be more complex devices, such as a whole functional unit or even an entire gene cluster (this is why "unitary" is used rather than "basic"). The distinction is in whether an assembly is referenced in the design (i.e., a composite part can be transformed into a unitary part by stripping associated assembly information).
-
Composite Part: A part that is designed as the composition of two or more other parts through an assembly. Note that samples of a composite part need not actually be produced by its designated assembly process: for example, the part might be implemented directly via synthesis, including the scars that would have been formed as part of assembly.
-
Assembled Part: A part, plus any 5' or 3' flanking scars, within the post-assembly context of a composite part.
-
Scar: A sequence that is produced by the combination of flanking sequences in an assembly.
-
-
Backbone: A DNA construct into which parts are intended to be inserted at one or more designated insertion sites, in order to meet the requirements of an assembly. Precisely one part can be inserted at any given insertion site. In many cases, a backbone will be a circular plasmid with precisly one insertion site, but other types of vector are possible as well, such as linear plasmids, viral replicons, or non-replicating flanking adapters.
-
Drop-Out Sequence: A portion of a backbone at an insertion site that is removed when a part is inserted at that site. Some backbones include drop-out parts while others do not.
-
Part Insert: A part, plus any 5' and 3' flanking sequences, that is placed into a designated insertion site of a backbone.
-
Part in Backbone: A backbone with at least one insertion site occupied by a part insert.
-
Part Extract: A part, plus any 5' or 3' flanking sequences, that has been extracted from a part in backbone as part of an assembly process. Note that the same extract can be produced from a backbone with flanking sequences and an insert without or a backbone without flanking sequences and an insert that includes them.
-
-
Assembly: A plan for combining a set of input parts in order to produce an output of either a single composite part or a library of composite parts. The inputs and output may or may not include backbones, depending on the specifics of the assembly. An assembly plan can be executed by appropriate laboratory protocols.
-
Device: Design for a functional mechanism in some biological context. A device may involve multiple parts, non-DNA elements, and interactions. For example, a small molecule sensor might involve two parts, a constitutive promoter/CDS combination and the promoter it results, as well as the protein product of the CDS, the small molecule inducer it binds to, and the production and regulation interactions. Note that functional characterization information pertains to devices, not parts. Note also that a single device might be instantiated by multiple different parts (e.g., a GFP expression device and several alternative encodings for the CDS implementing the device, all of which are expected to have equivalent expression in the intended biological context).
In the representational practices below, note that whenever an ontology term is specified, it should be taken to indicate either that term or any of its children. Any tools implementing these practices SHOULD NOT use strict equality tests that omit relationships between terms, even for ontology terms with no children, because children may be added to terms in future versions of the ontology.
Any part, unitary or composite, is represented by an SBOL Component with a type property of SBO:DNA. As per usual with SBOL, other functional information (e.g., promoter, CDS, complex function) is indicated by a Sequence Ontology term on the role property, and sub-structure of the part is indicated by its feature properties. A part MUST have precisely one sequence property.
Example: BBa_E1010 is a unitary part for an RFP reporter. The BBa_E1010 part, https://synbiohub.org/public/igem/BBa_E1010 is a Component of type SBO:DNA with role SO:CDS. It has asequence linking to https://synbiohub.org/public/igem/BBa_E1010_sequence, which is 706 bp going from start codon to stop codon plus a trailing barcode.
A composite part is represented by a Component that has its prov:wasGeneratedBy property set to a prov:Activity representing a RECOMMENDED assembly plan (see Assembly below).
Note that this implies that a composite part can be converted to a unitary part by removing its assembly plan (and vice versa, for cases in which that is possible).
Likewise, since prov:wasGeneratedBy can have multiple values, a composite part MAY record more than one possible alternative assembly plan.
An assembled part within a composite part SHOULD be represented by the composite Component including a feature that is a SubComponent with an instanceOf property linking to the part that has been included by assembly.
If the assembly has produced scars in the composite part, each scar SHOULD be indicated in the composite Component by a feature that is a SequenceFeature with role SO:restriction_enzyme_assembly_scar.
A scar SHOULD either meet or overlap an assembled part on both its 5' and 3' sides.
This SHOULD be indicated explicitly by Constraints with the meets or overlaps restriction and/or implicitly via the values of the features' hasLocation properties.
Example: BBa_K093005 is a composite part combining the BBa_B0034 RBS and BBa_E1010. It is a Component of type SBO:DNA with role SO:engineered_region and sequence BBa_K093005_sequence. BBa_B0034 is a SubComponent at bases 1-12 and BBa_E1010 is a SubComponent at bases 19-724.
Between them bases 13-18 should be marked with a SequenceFeature of type SO:restriction_enzyme_assembly_scar for the BioBrick RBS/CDS "TACTAG" scar.
The composite part should also have a prov:wasGeneratedBy indicating BioBrick assembly of BBa_B0034 and BBa_E1010.
Like a part, a backbone is represented by an SBOL Component with a type property of SBO:DNA. The role of a replicating backbone SHOULD be SO:vector_replicon or one of its children (e.g., SO:plasmid_vector).
A non-replicating backbone, such as used in linear fragments with flanking sequences for restriction, SHOULD instead use SO:engineered_region.
As with any Component, other information about the structure of the backbone is indicated by its feature properties.
Unlike a part, a backbone does not necessarily have to have a sequence.
This is because a backbone is not necessarily retained in the final construct and it is possible to use a backbone "blindly" like a reagent in the assembly process.
It is still RECOMMENDED that a backbone have a sequence, however, in order to facilitate better planning, debugging, and quality control.
Each insertion site on a backbone SHOULD be indicated with a feature that is a SequenceFeature with role SO:insertion_site.
For circular plasmids, an insertion site is commonly placed at the start/end of the sequence. In this case, the insertion site SHOULD be placed at 0 rather than at the end of the sequence.
A drop-out sequence in a backbone SHOULD be indicated with a feature with a role of SO:deletion. Note, however, that a drop-out sequence can be any sort of feature (e.g., the drop-out sequence might be a SubComponent for expressing a selection marker). When there is a drop-out sequence, the insertion site SHOULD be indicated at the start of the drop-out sequence (unless the drop-out sequence is at the end of a circular sequence).
A backbone MAY include flanking sequences for assembly at its insertion sites. Each flanking sequence SHOULD be indicated with a feature that is a SequenceFeature or SubComponent with role SO:restriction_enzyme_region.
Example: pSB1C3 is a Chloramphenicol-resistant high-copy plasmid used as a backbone for BioBrick assembly. It is a Component of type SBO:DNA with role SO:plasmid_vector. It should also have type SO:circular, since it is a circular plasmid. pSB1C3 includes BioBrick prefix and suffix flanking regions, that should be given role SO:restriction_enzyme_region.
The BioBrick prefix and suffix can be represented by a SubComponent that is an instanceOf BBa_G00000, and the suffix be a SubComponent that is an instanceOf BBa_G00001, which in turn should be marked with the the EcoRI, XbaI, SpeI, and PstI recognition and cutting sites should also be marked. For example in BBa_G00000, a SequenceFeature should mark bases 1-6 (gaattc) as with role SO:restriction_enzyme_recognition_site and another SequenceFeature should mark bases 2-5 (aatt) with the role SO:five_prime_sticky_end_restriction_enzyme_cleavage_site.
Example: A pair of flanking sequences for Type IIS assembly from fragments can be represented as a backbone Component of type SBO:DNA with role SO:engineered_region with three LocalSubComponent objects: one for the prefix, one for the insert, and one for the suffix. The prefix and suffix objects have role SO:restriction_enzyme_recognition_site and are linked to Sequence objects specifying their content. The insert has no information about its content, but is placed between the prefix and suffix by Constraint relations with the meets restriction.
A part inserted into a backbone is represented by a Component that includes both the part insert as a feature that is a SubComponent and the backbone as another SubComponent.
If the backbone includes the flanking sequences or no flanking sequences are needed for assembly, then in the SubComponent for the part, the value of its instanceOf property SHOULD be the Component for the part. The SubComponent should also indicate that it is the part insert by adding a role of SO:engineered_insert and a roleIntegration value of mergeRoles.
If the backbone does not include flanking sequences and they are needed for assembly, then the part insert SHOULD be a Component with role SO:engineered_insert that includes both the part and its flanking sequences as features. As with flanking sequences in the backbone, each flanking sequence SHOULD be indicated with a feature that is a SequenceFeature or SubComponent with a role in the SO:restriction_enzyme_region family.
More specifically, in the case of most digestion-based assembly methods, the roles will typically be SO: restriction_enzyme_five_prime_single_strand_overhang and SO: restriction_enzyme_three_prime_single_strand_overhang
If the insertion site is at the beginning or end of the backbone, then the relationship between the part insert and backbone MAY be indicated by Constraint relations with the meets or overlaps restriction.
Otherwise, the relationship between the part insert and backbone MUST be indicated via the values of their hasLocation properties.
Note that any insertion not at the end of a backbone implies that it will have a hasLocation for the portion of the backbone 5' of the insertion site and another hasLocation for the portion of the backbone 3' the insertion site.
If the part insert replaces a drop-out sequence in the backbone, then the drop-out sequence should be excluded from the part in backbone construct by excluding its range in the sourceLocation properties of the backbone SubComponent.
The representation of the part extract from a part in backbone is much like that of the part insert. If the part extract has no flanking sequences, then the part extract is simply the Component for the part.
If the part extract includes flanking sequences, then the part extract is represented by a Component that includes both the part and its flanking sequences as features. In some cases, the part extract is identical to the part insert, in which case the same Component can be used to represent both.
An assembly plan is represented using a prov:Activity object to link between the Component for a composite part and another Component that describes a sequence of reactions for producing the composite part from a set of input parts.
Specifically, the Component for the composite part SHOULD have a prov:wasGeneratedBy propert whose value is a prov:Activity object with two values of its type property, one being sbol:design and the other being the new term http://sbols.org/v3#assemblyPlan. The prov:Activity SHOULD also have a prov:qualifiedUsage property whose value is a prov:Usage with a prov:entity of a Component describing the assembly plan and a prov:hadRole of sbol:design.
The assembly plan component SHOULD have a SubComponent for the composite part and for each part used in its assembly.
Typically, however, these parts must be in backbone in order for the assembly to take place.
As such, the assembly plan Component SHOULD have a SubComponent for the composite part in backbone and a SubComponent for each part in backbone that is used in the assembly.
To indicate the relationship between each part and that part in backbone, the assembly plan SHOULD contain a Constraint with a type value of contains, the part in backbone as the value of subject, and the composite part produced as the value of object.
The assembly of the composite part in backbone starting other parts in backbone is described with a set of Interaction objects describing digestion to produce part extracts and ligation to produce one or more composite parts in backbone.
The assembly plan Component SHOULD also abstractly describe the assembly by an Interface that indicates each part used in the assembly as an input and the composite part as an output.
Specifically, a digestion step is represented by on Interaction of type SBO:cleavage. Each input vector or enzyme is indicated using a Participation object with a role property of value SBO:reactant and participant property whose value is the Feature for the vector or enzyme. Each part extract is also indicated using a Participation, except that the role value is SBO:product.
A ligation step is the same as a digestion step, except that the type of the interaction is SBO:conversion (NOTE: this should change to SBO:ligation on resolution of this SBO issue) and the part extracts are the reactants while the composite part is the product.
While many composite parts will be described with just one digestion/ligation stage, an assembly MAY have any number of digestion and ligation stages to produce the ultimately intended composite part.
An assembly plan MAY produce multiple composite parts outputs.
For example some intermediate products may be used in multiple composites or may themselves be considered ouputs.
In this case, the Interface will have multiple output values, one for each composite part that points to the assembly plan with its its prov:wasGeneratedBy.
An assembly can be validated by checking whether the features with SO:restriction_enzyme_region in the reactants and products of its interactions have the appropriate patterns for the enzymes.
Example: A Component describing an iGEM standard assembly producing to BBa_K093005 by combining BBa_B0034 and BBa_E1010 would contain a SubComponent for each of those three components. If the parts are all in the pSB1C3 vector, then the assembly plan will also include a SubComponent for "BBa_K033005 in pSB1C3" and a Constraint indicating that BBa_K033005 is contained in "BBa_K033005 in pSB1C3"), and similar for the other two.
The assembly is described with three Interaction objects. One has a type value of SBO:cleavage with BBa_B0034, EcoRI, and SpeI in Participation object with a role of SBO:reactant and a LocalSubComponent for "BBA_B0034 extract" as the SBO:product. Another with a type of SBO:cleavage has BBa_E1010, EcoRI, and XbaI with a role of SBO:reactant and a LocalSubComponent for "BBA_E1010 extract". The third Interaction has a type of SBO:conversion, a "BBA_B0034 extract" and "BBA_E1010 extract" with role of SBO:reactant and "BBa_K033005 in pSB1C3" with a role of SBO:product.
A device with precisely one instantiation via parts is represented by a Component object with a type of SBO:Functional Entity. Each part used is indicated via a SubComponent feature. Conversely, the device MUST NOT contain any other DNA-typed feature.
A device with multiple parts might end up with those parts being placed at different locations within a single plasmid or being placed on more than one different plasmid. In order to allow flexibility in how a device is used, the device Component SHOULD allow for either option unless there is a functional reason to constrain designs otherwise (e.g., some recombinase devices require parts to be on the same strand).
A device with multiple possible instantiations via parts is represented instead by a CombinatorialDerivation object, with the logical device being its template and the part options being indicated by VariableFeature objects. Note that if the device involves multiple parts and there are non-trivial constraints on part combinations, the representation may be complex. Accordingly, it is currently NOT RECOMMENDED that parts be combined together into devices in this manner.
Example: A device for mRFP1 expression could be implemented with a Component using two SubComponent features, one that is an instanceOf BBa_K093005, the other an instanceOf a Component for the mRFP1 protein. The relation between the two is expressed with an Interaction whose type has value SBO:genetic production, in which a ComponentReference for BBa_E1010 inside of BBa_K093005 is linked from a Participation with a role of SBO:template and the SubComponent for mRFP1 is linked from a Participation with a role of SBO:product. Additional measurement and modeling information about the effective translation rate in this device may be added to this Interaction via hasMeasure properties linking to om:Measurement, or by the hasModel property on the device. Information about mRFP1's fluorescence, decay rates, etc. would be kept on the Component for mRFP1.
A catalog of available parts in backbones, such as would be found in the iGEM distribution, can be implemented in terms of two Collection objects, one for all of the parts and one for all of the parts in backbone (possibly annotated with additional information, such as the location that the part in backbone should be able to be found in the distribution plates).
These part in backbone collection can be queried to find all of the part in backbones containing a given part P by searching the part in backbone collection for all Components that (recursively) contain a SubComponent whose instanceOf value is P.
Given a Collection of parts in backbone, the sequences needing to be synthesized into each backbone can be exported by searching each part in backbone Component for all of the SubComponent objects whose effective role in context contains a value of SO:engineered_insert.
Each such sequence needs to be synthesized and inserted at the adjacent insertion site on the backbone.
To check assembly compatibility for a part in backbone, one needs to check that the part has appropriate flanking sequences and that there are no illegal restriction sites.
A generic assembly template can be generated as a Component with digestion and ligation reactions, restriction enzyme recognition sites. The parts to be assembled, however, have only fully specified flanking sequence features with meet constraints on an undefined part feature. Likewise, the composite part has fully specified flanking sequence features and scars with meet constraints on undefined part features.
To check the part in backbone, create an assembly Component with SubComponent objects for the part the generic assembly template and add an equals constraint between the part in backbone and the undefined part in the template. Once this is done, attempt to calculate the sequence of the Component: if there is a conflict between the part in backbone and the flanking sequences of the generic assembly template, then the part in backbone does not have compatible flanking sequences.
Once the sequence has calculated, check for illegal restriction sites by scanning for all locations with sequences that match the recognition sites in the flanking sequences. Any such location that is not fully contained within a flanking sequence means that the part in backbone is not compatible with the assembly.
The fully computed assembly Component can serve as a validation record for assembly compatibility, if desired.
To design a composite part by assembling together parts, use a generic assembly template as when checking assembly compatibility. In this case, however, all of the parts to be assembled should be specified, not just one in the role being checked. Following the equality constraints should then allow the sequence of the composite part to be calculated.
Examples and use cases are embedded in the specification material given above. There are additional examples illustrated below using UML diagrams. In addition, Figure 1 shows how we have chosen to simplify the visual representation of some of the common patterns within these diagrams.
Figure 1: Common visual abstractions in UML diagrams, including (A) representation of a Constraint between two Features (in this case, SubComponents), (B) representation of a Constraint between two ComponentReferences to Features (in this case, SequenceFeatures), and (C) representation of two different types of Location.
Figure 2 and Figure 3 illustrate two ways of representing the composite part BBa_K093005, one with Locations and one with Constraints. BBa_K093005 is a product of assembling the RBS BBa_B0034 and CDS BBa_E1010 in accordance with the BioBrick RFC10 standard. A scar is formed between the parts as a result of their assembly.
Figure 2: Composite part example UML diagram with Locations.
Figure 3: Composite part example UML diagram with Constraints.
Figure 4 and Figure 5 illustrate two ways of representing the plasmid vector pSB1C3, one with Locations and one with Constraints. pSB1C3 includes the prefix and suffix of the BioBrick RFC 10 assembly standard (BBa_G00000 and BBa_G00001, respectively) and an insertion site between these flanking sequences.
Figure 4: Backbone example UML diagram with flanking sequences and Locations. The contents of BBa_G00000 are purposefully omitted for simplicity of presentation.
Figure 5: Backbone example UML diagram with flanking sequences and Constraints. The contents of BBa_G00000 are purposefully omitted for simplicity of presentation.
Figures 6, 7, 8, and 9 illustrate different ways of representing the plasmid vector pSB1C3 with a BBa_E1010 insert. Figure 6 and Figure 7 show how to represent BBa_E1010 in pSB1C3 with Locations and Constraints, respectively.
Figure 6: Part in backbone example UML diagram with flanking sequences in backbone and Locations. The contents of BBa_G00000 and BBa_G00001 are purposefully omitted for simplicity of presentation.
Figure 7: Part in backbone example UML diagram with flanking sequences in backbone and Constraints. The contents of BBa_G00000 and BBa_G00001 are purposefully omitted for simplicity of presentation.
By contrast, Figure 8 and Figure 9 show two ways to represent the assembly flanking sequences as part of the BBa_E101 insert instead of the pSB1C3 backbone. In Figure 8, these flanking sequences are the BioBrick RFC 10 prefix and suffix, but in Figure 9 they are overhangs resulting from cleaving the prefix and suffix with the restriction enzymes XbaI and PstI. Note that these overhangs are still contained by the RFC 10 prefix and suffix in the context of the overall "part in backbone" representation in Figure 9.
Figure 8: Part in backbone example UML diagram with flanking sequences in part insert and Locations. The contents of BBa_G00000 and BBa_G00001 are purposefully omitted for simplicity of presentation.
Figure 9: Part in backbone example UML diagram with flanking sequences in part insert/extract and Locations. The contents of BBa_G00000 and BBa_G00001 are purposefully omitted for simplicity of presentation.
Figure 10 and Figure 11 illustrate two ways of representing the result of extracting a BBa_E1010 insert from pSB1C3, one with Locations and one with Constraints. Extraction was done using the XbaI and PstI restriction enzymes.
Figure 10: Part extract example UML diagram with Locations.
Figure 11: Part extract example UML diagram with Constraints.
Figure 12 illustrates one way of representing the assembly of BBa_B0034 and BBa_E1010 into BBa_K093005 in accordance with the BioBrick RFC10 standard. Extraction of BBa_B0034 its part in backbone is represented with the EcoRI and SpeI restriction enzymes as participants, while extraction of BBa_E1010 is represented with the XbaI and PstI restriction enzymes as participants. The resulting part extracts with the appropriate overhangs are then represented as participants in their ligation into BBa_K093005. The positions of features within all parts and backbones are specified using Locations, and parts in backbones are represented with the flanking sequences for assembly placed in the pSB1C3 backbone as opposed to their engineering inserts.
Figure 12: Assembly example UML diagram with flanking sequences in backbone and Locations. The contents of BBa_G00000 and BBa_G00001 are purposefully omitted for simplicity of presentation. Objects are color-coded based on what they represent: orange is for the assembly plan, blue is for backbone, green is for part in backbone, yellow is for composite part, and pink is for part extract.
This SEP does not modify any existing definitions, and is thus backward compatible with SBOL 3.0.1
It would be desirable to link assembly more explicitly to protocols using the PAML protocol representation (http://bioprotocols.org). The recommendations in this SEP for representing assembly may be updated or followed by another SEP as that work becomes more mature.
Although representation of genomic integration is not explicitly within scope of this SEP, in many cases genomic integration can be represented in the same was as insertion of a part into a backbone. This can be done simply by substituting the genome for the backbone.
Specific recommendations regarding device measurements are out of scope of this SEP.
None at present
To the extent possible under law,
SBOL developers
has waived all copyright and related or neighboring rights to
SEP 055.
This work is published from:
United States.