OptBioDes: an optimal design of experiments (DoE) base package for synthetic biology

Basic Optimal DoE (Design of Experiments) routines for metabolic pathway design.

Modules:

• OptDes: main module of optimal DoE routines.
• doebase: routines to interface the current SynBioChem DoE input sheet template with the OptDes module.
• synbioParts: integration of OptDes with SBOL and DNA parts repositories.

The DoE specifications can be provided in tabular form (csv or xlsx) with the following columns:

• DoE position: Position in the genetic construct.
• DoE permutation class: When labeled as 1, it means that the genetic part can be rearranged in any other labeled position. This option is typically used for gene coding regions. It is assumed that each position forms a unit, i.e., rearrangements always take place by selecting one coding region from each labeled position.
• DoE designation: type of genetic part: origin | resistance | promoter | gene.
• Part number: identifier linking to the parts repository (SynBioHub, 'ICE`, etc.).

Install

Global installation

conda install -c conda-forge doebase

Local installation

git clone https://github.com/pablocarb/doebase.git
cd doebase
conda env create -f environment.yaml
conda activate doebase

Run

Basic call from within the library (see evaldes for details):

from doebase import makeDoeOptDes
factors, fnames, diagnostics = makeDoeOptDes(
      fact,
      size,
      seed=None,
      starts=1040,
      makeFullFactorial=False,
      RMSE=1,
      alpha=0.05,
      verbose=False,
      random=False
)

Input parameters:

• fact: a dictionary of sortable keys (the positions in the construct) containing doebase.spec objects with the following attributes:

  • positional: float

    1.0 or None depending if the genetic part can be rearranged

  • component: str

    origin | resistance | promoter | gene

  • levels: list

    levels of the genetic part (see Note)

Note

• Origin levels (plasmid copy numbers)

  [pl1, pl2, ... ]

• Resistance levels

  [res1, ... ]

• Promoter levels (and blanks)

  [prom1, prom2, ..., '-', '-', ... ]

• Gene levels (gene variants)

  [g1_1, g1_2, ... ]

• size: size of the library.

• seed: random seed.

• starts: number of the starts for the DoE algorithm.

• makeFullFactorial: make a full factorial rather than a library of the given size.

• RMSE, alpha: parameters of the DoE algorithm.

• verbose

• random: make a random rather than an optimal design.

Authors

• Pablo Carbonell