Millar group

The Centre for Engineering Biology is the University of Edinburgh's research organisation in interdisciplinary, Synthetic and Systems Biology, founded in 2023 as the successor to SynthSys, itself founded in 2012 as the successor to the Centre for Systems Biology at Edinburgh (CSBE). CSBE was newly founded in 2007. The Centre for Engineering Biology is a hub organisation that links across multiple, major funding awards, including an Engineering Biology Mission Hub, and EPSRC C-loop.

The dataset presents mathematical models of the gene regulatory network of the circadian clock, in the plant Arabidopsis thaliana. The work will be published as Urquiza-Garcia, Molina, Halliday and Millar, title "Abundant clock proteins point to missing molecular regulation in the plant circadian clock", in Molecular Systems Biology, 2025 doi 10.1038/s44320-025-00086-5.

Starting from the U2019.3 and U2020.3 models, this project rescales parameters to match protein levels that were predicted ...

Data, FMv2 model and simulations for the Chew et al. 2017 paper (bioRxiv https://doi.org/10.1101/105437 ), updated in 2022, mostly on the prr7 prr9 double mutant, with controls in lsf1 and prr7 single mutants. This is one of the outputs from the EU FP7 TiMet project, https://fairdomhub.org/projects/92.

This data archive was updated during submisson to the journal _in Silico _Plants in 2022, and a Snapshot was published. The updates are not changing the core data or the FMv2 model that has been ...

The dataset presents mathematical models of the gene regulatory network of the circadian clock, in the plant Arabidopsis thaliana. The work is published in Urquiza-Garcia and Millar, Testing the inferred transcription rates of a dynamic, gene network model in absolute units, In Silico Plants, 2021.

Starting from the P2011 model, this project corrects theoretical issues (EC steady state binding assumption) to form an intermediate model (first version U2017.1; published as U2019.1) model, rescales ...

Project to test effects of temperature cycles on expression of Arabidopsis florigen gene FT, and whether these are mediated by temperature-dependent leaf development or temperature-specific FT expression, or both. Re-used and extended Arabidopsis Framework Model v1 to address this question. Led by Hannah Kinmonth-Schultz in Kim and Imaizumi labs, collaborating with Millar lab.

Collection of models submitted to PLaSMo by Andrew Millar and automatically transferred to FAIRDOM Hub.

Project to test effects of natural compared to growth chamber 16:8 LD cycles, on expression of Arabidopsis flowering-time genes, and to define the genetic mechanisms and environmental triggers involved. Led by Young-Hun Song and Akane Kubota in the Imaizumi lab, with collaborators testing plants in parallel in Zurich and Edinburgh.

Click on Snapshot 2 to download data, models and analysis for Daniel Seaton et al. biorXiv 2017 https://doi.org/10.1101/182071 and Molecular Systems Biology, accepted Jan 2018, https://doi.org/10.15252/msb.20177962. Note that the published paper cannot be fully linked into this record as the DOI above was not live when we made the Research Object from this Investigation on FAIRDOMHub.

Data, models and simulations for the Chew et al. 2014 paper (PNAS, https://doi.org/10.1073/pnas.1410238111), using wild-type Arabidopsis ecotype Col-0 in standard 12hL:12hD growth conditions, compared to La(er) or Fei-0 accessions, or to plants overexpressing a micro RNA (miR156).

Files required for reproducibility of computational results. This include Docker file and python packages

Clock mutants for lhy-1/cca1-11, prr9/7, toc1, lux-4, elf3-1 were transformed with the genomic regions of the associated clock genes tagged with NanoLUC-3FLAG-10His. The tagged genomic constructs were transformed in the mutants using Agrobcterium ABI strain (kindly donated by Prof. Seth Davis University of York). T3 plants resistant to homozygous for BASTA resistance were phenotyped by luciferase imaging asessing period phenotype or plant architecture. Rescuing lines were then used for performing ...

Modelling and experiments for FMv2 components.

Simulations, parameter sensitivity analysis etc. for FMv2

Modelling and experiments for FMv2 as a whole; Testing Framework Model version 2 (FMv2)

Modelling and experiments for FMv2 as a whole; Testing Framework Model version 2 (FMv2)

Modelling and experiments for FMv2 as a whole; Testing Framework Model version 2 (FMv2)

Model simulations compared to experimental data from the literature (publications from Mizuno lab are linked), testing the FMv2.

Modelling and experiments for FMv2 as a whole; Testing Framework Model version 2 (FMv2)

Follow-up to the validation experiments on FMv2, testing candidate mechanisms for high malate and fumarate accumulation in the Arabidopsis double mutant prr7prr9 and its parent accession Col. New collaborations with the groups of Teresa Fitzpatrick and TiMet partner Samuel Zeeman.

Assorted files prepared during the publication process of the FMv2, its validation and testing, mostly focussed on the Arabidopsis double mutant prr7prr9 and its parent accession Col. Data from other studies that are described separately, and linked by Atribution to the File records under this Study.

Collection of models used in the introduction of absolute units into A. thaliana circadian clock models, with software resources and documentation. The models are inspired by P2011, published in Pokhilko et al 2012. The study contains Assays that link to the P2011 starting model and the models U2019.1 - .3 and U2020.1 - .3. Each model is shared as a human-readable file in the Antimony language and the associated, machine-readable SBML file, which was automatically generated using the SBML export ...

The P2011 model (linked in the Assay below) was rescaled to match TiMet RNA data in clock mutants from Flis et al. 2015, also linked here as separate mean and SD files. The raw TiMet data is available elsewhere on FAIRDOMHub.

Assays for model composition here, in order to share model files; potentially training and validation data in other Studies.

The models in this record were published in Flis et al. Royal Society Open Biology 2015. Their original IDs in the PlaSMo resource and IDs in Biomodels are given below. Please select files for download from the 'Related Items' list or the object tree/graph, below. 'SUBMITTED' is the original model version; 'SIMPLIFIED' removes SBML elements that were incompatible with SloppyCell software.

Original model: Arabidopsis clock model P2011.1.1 from Pokhilko et al. Mol Syst. Biol. 2012, ...

Metabolite analysis in clock mutants: Col-0 parent and mutants gi-201, toc1-101 and prr7prr9; WS parent and lhy/cca1 double mutant. Plants grown in Golm and harvested at End of Day and End of Night, , assays 22 major metabolites. More detail on TiMet wiki if required. Heteroscedastic t-tests to highlight most significant changes, without multiple-testing correction.

Leaf number at flowering data from literature for prr7 prr9 and Col wild-type plants under long photoperiods and short photoperiods

Seedling hypocotyl data from literature for prr7 prr9 and Col wild-type plants under various photoperiods

RNA timeseries data from TiMet for clock genes in prr7 prr9 and Col wild-type plants under 12L:12D cycle and LL

effects of 1% increase in each parameter, more detailed analysis of water content

correlations of starch mobilisation and fresh weight under single parameter changes

Combination of multiple sub-models to form Framework Model version 2

Comparison of simulated wild-type and prr7prr9 double mutant under 12L:12D cycles. Simulation with CVODE simulator via SBSI v1.5 framework.

Biomass, leaf number and metabolites in Col0 (WT), prr7, prr7prr9, and lsf1. Metabolite data from plants after 28 days of growth were analysed most (27 days 'end of night', 28 days 'end of day' and 'end of night'). The data file also includes data from 21 days of growth ('end of day' and 'end of night'), which is useful for comparison to early-flowering plants not tested here, such as the lhycca1 double mutant, that flower before 28 days, altering their physiology.

Biomass (fresh mass, dry mass), leaf numbers, leaf area, gas exchange and 12 metabolites in Col0 (WT), prr7prr9, and lsf1 (presented in the preprint/paper) and pgm (not analysed further).

Biomass (fresh mass, dry mass), leaf numbers, leaf area, gas exchange and 12 metabolites in Col0 (WT), prr7prr9, and pgm at days 29 and 35, presented in the preprint/publication, with most data also for Col and lhycca1 at days 21/22/23, not analysed further.

We suggest that the lower carbon assimilation rate measured in lhycca1 (see gas exchange data) might allow a calibirated simulation in the FMv2 model in future to incorporate the indirect effects of nightly carbon starvation in this genotype ...

Simulation data from FMv2 calibrated for experiment L&H2, an experiment run at 18.5C instead of the 20.5C of the replicate and related studies. The Excel file includes the mean and SD of the relevant experimental data, and the figure panels.

Transcript profiling by microarray in 4, 6, 8, 12 and 18 h photoperiods, originally published in Flis et al, 2016, Photoperiod-dependent changes in the phase of core clock transcripts and global transcriptional outputs at dawn and dusk in Arabidopsis. doi: 10.1111/pce.12754.

Plant material The same plant material used for transcriptome analysis in (Flis et al., 2016) was the basis of our proteome study. Briefly, Arabidopsis thaliana Col-0 plants were grown on GS 90 soil mixed in a ratio 2:1 (v/v) with vermiculite. Plants were grown for 1 week in a 16 h light (250 μmol m−2 s−1, 20 °C)/8 h dark (6 °C) regime followed by an 8 h light (160 μmol m−2 s−1, 20 °C)/16 h dark (16 °C) regime for one week. Plants were then replanted with five seedlings per pot, transferred for ...

These Python scripts define and simulate the translational coincidence model. This model takes measured transcript dynamics (Blasing et al, 2005) in 12L:12D, measured synthesis rates of protein in light compared to dark (Pal et al, 2013), and outputs predicted changes in protein abundance between short (6h) and long (18h) photoperiods. These are compared to the photoperiod proteomics dataset we generated.

RNA timeseries data for Arabidopsis Col wild-type plants and clock mutants, as separate mean and SD files. The raw data is available on BioDare.ed.ac.uk, and is linked as 'Attribution' from elsewhere on FAIRDOMHub.

The starting models are included here in their original forms, the P2011 model as an SBML L3V1 model file, and the KF2014 model of Fogelmark et al. shared as SBML; both prepared by Uriel Urquiza.

Proteomics data for N15 incorporation into protein in Ostreococcus grown in 12L:12D light:dark cycles.

https://doi.org/10.1105/tpc.15.00737

https://doi.org/10.1038%2Fnplants.2017.87

The file contains the matrices that come from the MCMH sampling during the inference process from PBMs

The file contains the matrices that come from the MCMH sampling during the inference process from PBMs

https://doi.org/10.1073/pnas.1316278111

https://doi.org/10.1016%2Fj.cub.2010.12.021

Transformation of RLU into absolute units using a calibration curve of recombinant MBP-NanoLUC-3FLAG-10His

Plot of linear regresion of calibration curve for inferring number of molecules from NanoLUC biolumienescence in plant extracts

4 seeds of stable NanoLUC T3 homozygous lines for LHY, PRR7, TOC1, and ELF3 were seeded in 96-well flat white plate that contained 150 µl of ROBUST media and stratified for 2 days at 4ºC. Then plates were incubated in a 2 hours pulse of white light given and transferred to 22 hours darkness at 21 ºC. Then transferred 12L:12D photoperiod for 10 days. On day 10, 50 µl of 1:50 Furimazine:0.05% Triton X-100 added to each well for tracking NanoLUC bioluminescence. A) Measurements using a Tristar plate ...

4 seeds of stable NanoLUC T3 homozygous lines for LHY, TOC1 and ELF3 were seeded in 96-well flat white plate that contained 150 µl of ROBUST media and stratified for 2 days at 4ºC. Then a 2 hours pulse of white light given and transferred to 22 hours darkness at 21 ºC. Then transferred 12L:12D photoperiod for 10 days at which 50 µl of 1:50 Furimazine:0.05% Triton X-100 added to each well for tracking NanoLUC bioluminescence. Measurements using a Tristar plate reader were performed automatically ...

Small data base of clock proteins profiles obtained with western blots by several authors of A. thaliana collected from the literature. This data can be fed into simple models for making prediction of abosute number of protein when combined with RNA data in absolute units. For example the TiMet data set

"Samples of plants were collected in pre-weighed 2 ml microfuge tubes (safelock, Eppendorf) with 5 mm stainless steel grinding balls, and flash frozen in liquid nitrogen. The tissue was ground twice at 30Hz for 1 min in a Tissue Lyser (Qiagen). The samples were flash frozen between grinding steps, then placed on ice and 150 μl of BSII buffer was added to protect the samples from proteolysis, without phosphatase inhibitors (Huang et al. 2016). The tube was weighed and further BSII buffer added to ...

"Plates inoculated with Col-0 seed were grown under the same photoperiod conditions to the plants to be analysed. Plant tissue was harvested, making aliquots of 0.4 gFW. MBP-NL3F10H protein was prepared by the method described by Urquiza-Garcia U. and Millar A.J. in Plant Methods 2019. and then quantified by the linearized Bradford assay protocol using both Bovine serum albumin BSA and Ovoalbumin as standards (Ernst & Zor 2010). Then aliquots spiked with purified enzyme to generate a curve ...

This time series were obtained from the literature by perroming rough quantiftification from western blot images. In some cases the data was quantified by the authors and graphs were provided in the publications. In this case we used ImageJ or https://automeris.io/WebPlotDigitizer/

This files contains the predicions generated using a simple model of translation, described in the manuscript. This synthetic data was used to rescale U219.3 resulting in U2019.4 and U2020.3 into U2020.4. The .4 models are only resceled for the mass scale of protein and still present the dynamics of the .3 version

Analysis of carbon metabolites in clock mutants by Anna Flis and Ronan Sulpice, Mark Stitt lab

Excel spreadsheet with data and simulations used to prepare figures for publication, see Metadata sheet for conditions. Data Fresh (not dry) rosette leaf biomass, measured in samples of 5 plants each on multiple days, as mean and SD; Simulation outputs from FMv2 for Col Wild Type plants, and two simulations for prr7prr9 where the mutation affects only starch degradation or both starch degradation and malate/fumarate store mobilisation.

Starch levels in carbon units (not C6) measured on on days ...

Excel workbook with included Read.Me sheet, including FW and DW biomass data derived from files linked elsewhere; a compilation of the rosette area and gas exchange data for every plant measured of the Col, lsf1 and prr7prr9 genotypes; statistical analysis across the experiments; and charts of the compiled data, some of which are presented as figure panels in the 2022 versions.

Parameters rescaled and scaling factors set to 1

This is the scaled version of U2020.4 in sbml file. It already contains the scaling factors

Paramters rescaled and scaling factors set to 1

Paramteres rescaled and scaling factors set to 1

Parameters rescaled and scaling factors set to 1

Derived from U2019.3 from Testing the inferred rate of dynamic, gene regulatory network in absolute units

Sbml version of U2019.4 with reacaling factors values already incoporated in the model. This was generated autmatically using tellurium python package

This file was derived from U2020.3 by introducing the scalig factors in the required locations in the model. This files is used then for numerically rescaling the model for matching synthetic protein data.

A model of the circadian regulation of starch turnover, as published in Seaton, Ebenhoeh, Millar, Pokhilko, "Regulatory principles and experimental approaches to the circadian control of starch turnover", J. Roy. Soc. Interface, 2013. This model is referred to as "Model Variant 2". The other model variants are all available from www.plasmo.ed.ac.uk as stated in the publication. Note that the 'P2011' circadian clock model was modified for this publication (as described), in order to replicate the ...

Matlab model (could not be represented in SBML) from publication with abstract: Clock-regulated pathways coordinate the response of many developmental processes to changes in photoperiod and temperature. We model two of the best-understood clock output pathways in Arabidopsis, which control key regulators of flowering and elongation growth. In flowering, the model predicted regulatory links from the clock to CYCLING DOF FACTOR 1 (CDF1) and FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) transcription. ...

Model derived from U2019.2, fitted to TiMet data mutants data set. Fixed parameters are scaling factors, COP1 and cP parameters. The rest of the parameters were left optimisable. The networks used in the fitting include WT, lhycca1, prr79, toc1, gi and ztl. The ztl network was only used for fixing the period in this mutant. Then final parameter values for transcription rated were obtained by taking the product of scaling factor and either transcription or translation, the latter required for ...

Model derived from U2019.1 in which the transcription rates were rescaled to match the scale of TiMet data set for absolute units of RNA concentration. The gmX scaling parameters in the model were fitted numerically. This model has equivalent dynamics to P2011.1.2.

Model derived from U2020.2, fitted to the TiMet RNA data for wild-type and clock mutants. Fixed parameters are scaling factors, COP1 and cP parameters. The rest of the parameters were left optimisable. The networks used in the fitting include WT, lhycca1, prr79, toc1, gi and ztl. The ztl network was only used for fixing the period in this mutant. Then final parameter values for transcription rates were obtained by taking the product of scaling factor and either transcription or translation, the ...

Model derived from U2020.1 by fitting the scaling factors for matching TiMet data set for wild-type and clock mutants, in absolute units.

Model derived from U2019.1, in which the way the PRR genes are regulated is modified. Repression mechanism introduced Instead of activation between the PRRs for producing the wave of expression. This is inspired in the result of three models P2012, F2014 and F2016. P2012 introduced TOC1 repression in earlier genes relative to its expression. F2014 introduced also the backward repression of PRR9 |-- PRR7 |--- PRR5, TOC1. However little attention was given to why there is a sharper expression ...

Model written in Antimony human-readable language and then translate into SBML using Tellurium

Model written in Antimony human-readable language, Model used in Pokhilko et al 2012

autogenerated equation listing from the SBML of U2020.3, as a .PDF file

autogenerated equation listing from the SBML of U2019.3, as a .PDF file

Outline report of joint research conducted during MSBnet-funded visit of Sanu Shameer to Millar lab

Construct fof TOC1 for comparision with NanoLUC

Binary vector used for transforming with Agro ABI the elf3-2 CCA1p:LUC for rescuing elf3-2 mutation. In this particular case the plants selected were based on hypocotyl length. Based on the data of LUX in which we observed that rhytmicity was rescued in all lines, hypocotyl elongation varied between lines. Therefore, we used hypocotyl length for assesing complementation

Donor plasmid for Gatway cloning

https://benchling.com/s/02xiS4zk?m=slm-xxhynNMU6VkdMdqdfpFr

Donor plasmid for Gatway cloning

https://benchling.com/s/3OtoUun0?m=slm-pL4lmqV9bCWOW3WvoSFL

Binary vector used for transforming with Agro ABI the cca1-1/lhy-1p:LUC for rescuing cca1-1 mutation.

https://benchling.com/s/Zu0nJONs?m=slm-5zxnih8JIOujB9nYIG8E