ON : 5’ GGGACCCAGUAGCCAAAAGGGUACUUGUUGAGUAGAGUGUGAGCUCCGUAACUAGUCGCGUCUUC
GGACGCACUGAAUGAAAUGGUGAAGGACGGGUCCAGUGCCUAGAUGUAGCUAAGCUACAUCUGGCUACUGGGU3’

Figure 1. NUPACK A) FASTmiR-222 OFF state B) FASTmiR-222 ON state

Figure 2. A)RNAfold FASTmiR-222 OFF state B)RNAcofold FASTmiR-222 ON state

FASTmiR sensors with different split positions specific for miR-146a, miR-27a-5p, miR-30c and other miR-222 were similarly made with the same strategy in mind. Here are the other designed sequences.

FASTmiR-222 FASTmiR-27a-5p FASTmiR-30c-1 FASTmiR-146a-3p

Figure 1. A) FASTmiR-222-D1 OFF state B) FASTmiR-222-D1 ON state

Figure 1. A) FASTmiR-27a-5p-D1 OFF state B) FASTmiR-27a-5p-D1 ON state

Figure 1. A) FASTmiR-27a-5p-D2 OFF state B) FASTmiR-27a-5p-D2 ON state

Figure 1. A) FASTmiR-27a-5p-D3 OFF state B) FASTmiR-27a-5p-D3 ON state

❮❯

❮❯

❮❯

Figure 1. A) FASTmiR-30c-1-D1 OFF state B) FASTmiR-30c-1-D1 ON state

Figure 2. A) FASTmiR-30c-1-D2 OFF state B) FASTmiR-30c-1-D2 ON state

Figure 3. A) FASTmiR-30c-1-D3 OFF state B) FASTmiR-30c-1-D3 ON state

❮❯

❮❯

❮❯

Figure 1. A) FASTmiR-146a-3p-D1 OFF state B) FASTmiR-146a-3p-D1 ON state

Figure 2. A) FASTmiR-146a-3p-D2 OFF state B) FASTmiR-146a-3p-D2 ON state

Figure 3. A) FASTmiR-146a-3p-D3 OFF state B) FASTmiR-146a-3p-D2 ON state

❮❯

❮❯

❮❯

Dual Aptamer Design

STRATEGY

1. Choose an aptamer (single-stranded RNA or DNA) that has the following:
   • • high affinity (low dissociation constants; aptamers can bind to even low amounts of biomarkers)
   • • high specificity (prevents cross-reactivity with other molecules present in the sample) and
   • • high selectivity (binds maximally to the desired biomarker and folds lesser to non-target molecules)
2. Using literature, obtain the potential target-binding segments within the aptamer
3. Aptamer complementary sequence: 
   • • Analyse the secondary structure of aptamer using DuplexFold, UNAFold, NUPACK, etc. [1-3]
   • • Design different sequences, blocking whole binding sites, regions adjacent to binding sites or overlapping binding and non-binding regions of the aptamer. If these sites are not known or aptamer is not well-characterised, go by hit and trial method by analysing secondary structures and minimum free energy values.
   • • Vary the length of the complementary sequences. Some mismatched bases can also be incorporated.
4. Experimentally, with different concentrations of biomarkers, test the displacement response to the aptamer’s conformational change when the biomarker is recognised.

PARAMETERS

1. The biomarker and its aptamer should be more stable than the aptamer-trigger complex. Only then all the triggers can get displaced, ensuring that the reaction goes in one direction.
2. The affinity of the aptamer-trigger should be strong enough to remain in annealed form in the absence of a biomarker and weak enough so that trigger gets easily displaced in the presence of a biomarker. The system's sensitivity will depend on the feasibility of biomarker-induced trigger displacement.
3. The target should be designed so that all the displaced triggers bind to the target after the sequential addition of the amplification module. This can be achieved by creating the target sequence entirely complementary to the trigger sequence.

APPLICATION

We designed two modules to detect two of our biomarkers, testosterone and C-reactive protein (CRP). For proof-of-concept, we used quinine as it has well-characterised aptamer (cocaine binding aptamer; MN4) and progesterone. Progesterone is a steroidal hormone and quinine is an alkaloid.

• • Testosterone
• • Progesterone
• • Quinine
• • C-reactive protein 

Aptamers that were chosen for the above molecules

Table 2. Information about the chosen aptamers

*at 140 nM NaCl concentration

EXAMPLE

Testosterone aptamer sequence (T6)

The dissociation constant, reported in the literature, lies in the low nanomolar range and was calculated using apta-PCR affinity assay (APAA) [10].

5’-TAGGGAAGAGAAGGACATATGATTCCTGTCGAATTCAAATCGAACTAGCCTCATCTCAGCTCGTTGACTAGTACATGACCACTTGA-3’

The sequence motifs for testosterone aptamer T6 can form the potential binding site for the target molecule. 

Table 3. Represents data obtained from NUPACK for the trigger sequences designed for testosterone aptamer T6 (Minimum free energy at 37° C is -3.88 kcal/mol; K_D^# = 1.8 ± 0.3 nM)

Testesterone Progesterone Quinine CRP

Figure 1. A) T6 Aptamer structure B) Trigger 1 design C) Target 1 design D) T6 Aptamer-Trigger complex

Figure 2. A) T6 Aptamer structure B) Trigger 2 design C) Target 2 design D) T6 Aptamer-Trigger complex

Figure 3. A) T6 Aptamer structure B) Trigger 3 design C) Target 3 design D) T6 Aptamer-Trigger complex

Slide 4

❮❯

❮❯

❮❯

Figure 1. A) P4G13 Aptamer structure B) Trigger 1 design C) Target 1 design D) P4G13 Aptamer-Trigger complex

Figure 2. A) P4G13 Aptamer structure B) Trigger 2 design C) Target 2 design D) P4G13 Aptamer-Trigger complex

Figure 3. A) P4G13 Aptamer structure B) Trigger 3 design C) Target 3 design D) P4G13 Aptamer-Trigger complex

Figure 4. A) P4G13 Aptamer structure B) Trigger 4 design C) Target 4 design D) P4G13 Aptamer-Trigger complex

Figure 5. A) P4G13 Aptamer structure B) Trigger 5 design C) Target 5 design D) P4G13 Aptamer-Trigger complex

❮❯

❮❯

❮❯

❮❯

❮❯

Figure 1. A) MN4 Aptamer structure B) Trigger 1 design C) Target 1 design D) MN4 Aptamer-Trigger complex

Figure 2. A) MN4 Aptamer structure B) Trigger 2 design C) Target 2 design D) MN4 Aptamer-Trigger complex

Figure 3. A) MN4 Aptamer structure B) Trigger 3 design C) Target 3 design D) MN4 Aptamer-Trigger complex

Figure 4. A) MN4 Aptamer structure B) Trigger 4 design C) Target 4 design D) MN4 Aptamer-Trigger complex

Figure 5. A) MN4 Aptamer structure B) Trigger 5 design C) Target 5 design D) MN4 Aptamer-Trigger complex

Figure 6. A) MN4 Aptamer structure B) Trigger 6 design C) Target 6 design D) MN4 Aptamer-Trigger complex

Figure 7. A) MN4 Aptamer structure B) Trigger 7 design C) Target 7 design D) MN4 Aptamer-Trigger complex

❮❯

❮❯

❮❯

❮❯

❮❯

❮❯

❮❯

Figure 1. A) HsaCRP Aptamer structure B) Trigger 1 design C) Target 1 design D) HsaCRP Aptamer-Trigger complex

Figure 2. A) HsaCRP Aptamer structure B) Trigger 2 design C) Target 2 design D) HsaCRP Aptamer-Trigger complex

Figure 3. A) HsaCRP Aptamer structure B) Trigger 3 design C) Target 3 design D) HsaCRP Aptamer-Trigger complex

Figure 4. A) HsaCRP Aptamer structure B) Trigger 4 design C) Target 4 design D) HsaCRP Aptamer-Trigger complex

Figure 5. A) HsaCRP Aptamer structure B) Trigger 5 design C) Target 5 design D) HsaCRP Aptamer-Trigger complex

Figure 6. A) HsaCRP Aptamer structure B) Trigger 6 design C) Target 6 design D) HsaCRP Aptamer-Trigger complex

❮❯

❮❯

❮❯

❮❯

❮❯

❮❯

PARTS MODIFICATION

Towards the modification of parts from the Registry of Standard Biological Parts (RSBP), we aimed to design investigations that would enhance the fluorescence intensity of Spinach2.1, a light-up aptamer [17]. In order to achieve this, we altered the sequence by introducing mutations (single, double), inversions, and deletions at different regions of the Spinach2.1 aptamer sequence to increase its relative fluorescence intensity than the wild-type. We employed two approaches in which, firstly, the sequence in the tetraloop (UUCG or TTCG) was mutated and inversion near this tetraloop sequence [10, 12]. Secondly, point mutations (A12G and U86C) were made at different places in the Spinach2.1 sequence. These mutations were chosen by a thorough inspection of the Spinach2.1 3D structure (PDB entry: 4TS2) with the help of Dr Hussain Bhukya, IISER Tirupati, and Dr Harikrishna S, Senior Scientist at Syngene.

All possible single-point mutations (SPMs) were made in the tetraloop using combinatorics and inversions. The free energy of the thermodynamic ensembles, frequency of the minimum free energy (MFE) structure, and the ensemble diversity were predicted using RNAfold. Based on these predictions, we chose a few sequences that displayed the best parameters. The mutated sequences and their corresponding RNAfold structures are given Figure 1.

Mutations and Inversion: 

The free energy of the thermodynamic ensemble are predicted to be -34.44 kcal/mol, -34.44 kcal/mol, -34.80 kcal/mol and -35.50 kcal/mol for the sequences (modified Spinach2.1) of Spinach2.2, Spinach2.3, Spinach2.4 and Spinach2.5 respectively.

Figure 1. Sequence and structures of the RNA Aptamers. The DNA sequence and the RNAfold structures of Spinach2 and Spinach2.1 are adopted from the parts registry made by DTU, Denmark. The Spinach2 sequences highlighted in bold are the aptamer sequence, and those flanking are the tRNA scaffold sequence, tRNALys3. The yellow colored box highlights the tetraloop sequence (colored red) and the bases adjacent (colored blue) to the tetraloop. The bases colored green are mutated from A/T and T/A to give a modified Spinach2 version, Spinach2.1. The other Spinach aptamer versions are derived from the Spinach2.1 sequence by base mutations in the tetraloop and the region adjacent to it (highlighted in the yellow colored boxes) and keeping the rest of the sequence unchanged. These mutants correspond to Spinach2.2, Spinach2.3, Spinach2.4 and Spinach2.5. The mutations in the DNA sequence are displayed on the RNAfold structures of the RNA aptamers (Spinach). The tetraloop region is indicated by a box colored blue for Spinach2 and the other mutations are circled in blue for other modified Spinach aptamers.

In silico Analysis: Molecular Dynamics (MD) Simulation of RNA-ligand interactions

Inspired by the fluorescence property of the DFHBI bound Spinach2.1 aptamer complex, we aimed to enhance the detection limits of this aptamer by modifying its sequence. To check whether the proposed modifications would enhance the fluorescence intensity, we performed in-silico studies by simulating the native DFHBI-Spinach aptamer complex (PDB entry: 4TS2) to validate the forcefield and other parameters for running Umbrella Sampling MD in Gromacs [13]. Later, the modified Spinach structure was generated in X3DNA[16] where the base mutations, A12G and U86C, are incorporated in the 4TS2 to give the DFHBI bound Spinach aptamer structure, Spinach2.6. Both the native and modified Spinach structures were subjected to identical simulation parameters in order to estimate the binding energy (ΔGbind) using the potential of mean force (PMF). The PMF was extracted from a series of umbrella sampling simulations and these values are anticipated to provide insights into the affinity of DFHBI binding to the G-quadruplex, a part of the aptamer structure [11,13,14].

Figure 2. Structures of DFHBI-Spinach RNA aptamer complex and Spinach2.6 complex generated in X3DNA are displayed in the left and right panels, respectively. The bases highlighted in spacefill model are the mutation points. The RNA sequence corresponding to the structures are displayed below and the bases mutated are shown in blue. 

GROMACS simulations

The input file for both the structures were generated using CHARMM-GUI, where the RNA aptamer and ligand (DFHBI) complex reside in the water (TIP3P water system) box consisting of the defined periodic boundaries. The Amber99-OL3 RNA force field was used to perform the energy minimization and equilibration (frame rate, 125000 ns steps x 2 fs) by normal volume and temperature (NVT) method for 250 ps to get the equilibrated structures of both the native and modified RNA aptamers. Steer MD simulations were performed on these equilibrated complex structures where an imaginary spring force was used to pull the ligand out of the binding site. The RNA was restrained in the course of pulling the ligand out of the G-quadraplex structure, where the trajectories of this course were plotted. Later, a window spacing was performed to generate configurations to help visualize the unbinding and rebinding . This was followed by adding-up the configuration free energies to obtain the final histogram and the total binding energy. 

Figure 3. Flowchart for GROMACS input file preparation.

The energy minimization results were plotted using the steepest descent algorithm and it was observed that the native structure converged in almost 200 steps (Figure 4a) while for Spinach2.6 converged in nearly 180 steps (Figure 4d). An NVT equilibration was performed and trajectory files were generated to visualise the simulation. It was found that the system reached equilibrium at less than 20 ps (Figure 4b, e). This equilibrated system was subjected to the Steer MD simulations to dislocate the ligand from the G-quadruplex binding site with spring force 10000 KJ/mol/nm², see the Figure 4c, d. The force applied was able to pull the ligand apart from the binding site of the RNA aptamer as seen in the animation below. Overall, the energy minimization plot suggests that the Spinach2.6 complex structure is relatively stable when compared to the native aptamer structure. Moreover, the Steer MD plots for both these complexes are significantly different suggesting the binding affinities of the ligand to RNA aptamer structure are different. However, additional simulations and experimental confirmations need to performed to strengthen this observation.

Figure 4.Results of MD simulations of native and modified RNA aptamers:a, b) Energy minimization and equilibration plots of native (4TS2) complexes respectively. c) Steer MD simulation plot for the native structure. d, e) Energy minimization and equilibration plots of modified Spinach RNA aptamer, Spinach2.6 complex structure respectively and f) Steer MD simulation plot for the Spinach2.6 complex structure

Steer MD trajectory visualisation was done using pull.gro and pull.xtc files.

REACTION KINETICS

One of the primary goals of modelling is to understand the behaviour of the system and check whether the system is behaving as expected. In this section, we shall be reviewing the detection systems by building rate law models and analysing the results through graphs.

miRNA Detection Protein and Hormone Detection

Introduction

The concentration of miRNA in our sample is quite low hence we proposed to amplify our miRNA through a process called miRPA [18] (Check Blueprint to know more about miRPA). Using the data from this model, we could predict the concentration of miRNA and miDNA that has been amplified for a given time and resources. Based on the amplified concentration, the amount of FASTmiR aptamers required was calculated and used to find out the final concentration of FASTmiR aptamer that is bound to the dye. 

Main goals

1. Determine the concentration of the amplified miRNA(and miDNA) in PCOS and non PCOS people and find the fold change.
2. Predict the concentration of FASTmiR aptamers required.
3. Find out the total concentration of FASTmiR sequences that is bound to the dye.

Assumptions

1. Most reactions have been considered to be irreversible since forward rate of reaction >> reverse rate
2. Some parametric values have been assumed because some reactions are specific to our system.

Reaction System

Figure 1. Flowchart Representing the Reaction System

Reactions

Equations

Parameters

Parameter	Description	Value	Units	Source
k1	Rate of probe 1 binding to miRNA	106	M-1s -1	Assumed
k2	Rate of probe 2 binding to complex 1 (miRNA - probe 1)	106	M-1s-1	Assumed
Vmax3	Rate of ligase activity (kcat*enzyme concentration)	2.4x10-6	Ms-1	Assumed
kM3	Michelis Menten constant of ligase	10-9	s-1	Assumed
k4	Rate of forward primer binding to duplex	106	M-1s-1	Assumed
Vmax5	Rate of Strand displacing polymerase activity	2.4x10-6	Ms-1	Assumed
kM5	Michaelis Menten constant of strand displacing polymerase	43x10-6	s-1	Assumed
k6	Rate of forward primer and recombinase	4.6x10-3	s-1	[19]
k7	Rate of forward primer-recombinase and miDNA	108	M-1s-1	[19]
k8	miRNA- FASTmiR aptamer Binding rate	106	M-1s-1	Assumed
k9	FASTmiRoff-dye Binding rate	106	M-1s-1	Assumed

Tunable components:

By tuning the following components

1. Concentration of miRNA 
2. Concentration of Probes
3. Concentration of Forward primer
4. Concentration of recombinase

Simulation Results

Figure 2: a) Rate of formation of miDNA for PCOS (2x10^-14 M) and Normal(1x10^-14 M) condition. b) Rate of formation of FASTmiRon aptamers for PCOS (9nM) vs Normal (4nM) condition

Results and Inferences:

Using these graphs we were able to find that after 30 minutes of amplification, which is the standard time of reaction of miRPA, we could reach a considerable concentration of miDNA. The miRNA is amplified to approximately 4nM and 9nM respectively for normal and PCOS case. It was noted that the fold change was nearly maintained at the same level before amplification. We could also find the time at which the maximum concentration of FASTmiR - dye complex is reached. This graph can be used to fix a threshold for PCOS and non-PCOS conditions. The plateauing of the concentration of the FASTmiR on-state that happens at the same concentration level, indicates that there is a one to one correspondence between miRNA concentration and the FASTmiR-dye complex concentration which verifies our biosensor working. Initially as the concentration of the miRNA increases there is a drastic increase in the concentration of the FASTmiR on-state. After a timepoint, it reaches a saturation point, after which, there is no further increase in the concentration of the FASTmiR on-state, which indicates the equilibrium state where most of the concentration of miDNA and miRNA is detected

Improvements:

In the model, we have assumed that non-specific interactions between miRNA and FASTmiR to be negligible. Since it can have a significant effect in the real system our model can be updated based on experimental data about the Kd values specific to our biosensors.

CLICK TO READ PROTEIN AND HORMONE DETECTION

Introduction

Proteins and hormones detection was devised by adapting a dual aptamer approach which involves in vitro transcription (Click here for more info.). A model with rate law equations of the system was developed by taking reference from literature.

Main goals

1. Determine the rate of formation of light up aptamers for different biomarkers.
2. Determine the time taken to achieve required concentration of light up aptamers.
3. Optimise the model by changing the parameters.
4. Build a Fluorescence vs time graph based on Calibration curve for all 3 biomarkers.

Assumptions

1. Enzymes are highly efficient even at low concentrations.
2. Dye concentration is present in excess.
3. The concentration of all resources needed for in-vitro transcription have been together named as Transcription Resources (TsR).
4. Some parametric values have been assumed due to lack of available studies.

Reaction System

Figure 1. Flowchart representing reaction system

Reactions

Equations

Parameters

Parameter	Description	Value	Units	Source
k1f	Rate of ATBC production	1	M-1s-1	Assumed
k1r	Rate of ATBC degradation	109	min-1	Assumed
k2	Rate of trigger production	2	min-1	Assumed
k3	Rate of Half duplex production	106	Ms-1	[20]
Kt1	Extension rate of phi29 polymerase	10-7	Mmin-1	Assumed
KM1	apparent nucleotide binding constant	31x10-6	M	[21]
Kt2	Transcriptional rate of T7 polymerase	18.2x10-9	Mmin-1	[21]
KM22	Michaelis constant	8.5x10-9	M	[21]
Kc	Rate of depletion of Transcriptional resources	0.011	min-1	[21]
k4f	Rate of binding of Dye	3.72x106	M-1min	[21]
k4r	Rate of Unbinding of Dye	1.44	min-1	[21]

Simulation Results

Figure 2. a) Rate of formation of Broccoli for different concentrations of biomarker for Normal (25nM) and PCOS (62nM). b) Theoretical graph for fluorescence vs time based on calibration curve.

Figure 3. a) Rate of formation of Broccoli for different concentrations of biomarker for Normal (3.8 nM) and PCOS (42.4 nM). b) Theoretical graph for fluorescence vs time based on calibration curve.

Figure 4. a) Rate of formation of Broccoli for different concentrations of biomarker for Normal(8.6 pM) and PCOS(19.7 pM). b) Theoretical graph for fluorescence vs time based on calibration curve.

Interpretation 

From the graphs, we see that the higher the biomarker concentration, the faster the rate at which Broccoli aptamers - dye complexes are formed. It can also be seen that fluorescence takes a long time to reach a significance level. It approximately takes 3.5 hours to reach an equilibrium. We can then fix a threshold for PCOS and non-PCOS conditions from the fluorescence vs time graph.

Improvement

We have built a simplistic model with lesser variables. We could check the sensitivity of the model towards factors like temperature. Actual Kd values for our aptamer system can be experimentally found and used in our models to improve accuracy of simulations

CLICK TO READ miRNA Detection

BIOMARKER ANALYSIS

A person might not be positive for all the biomarkers. Hence we needed to identify the combination of biomarkers that gives the most precise results in the given samples. Also, since we are detecting our biomarkers in a concentration-dependent manner, we needed to set a threshold to distinguish between PCOS and non-PCOS people. These results will be used to analyse the results by the Arduino (Click Hardware to know more about Arduino). We consulted with Dr Shibdas Banerjee and decided to take a machine-learning approach to classify PCOS and non-PCOS. 

Main goals

1. Build a machine-learning model for our biomarkers using a testing set.
2. Find out the best hyperplane which separates the two classes of our data. 
3. Validate it using the test set.
4. Predict the best combinations and thresholds for the biomarkers

We aimed to create a dataset for the biomarkers and build the model. However, due to a lack of experimental data, we envision a build a model in the future. We believe this model allows us to test the specificity and sensitivity of the biomarker combinations for detecting PCOS.

REFERENCES

1. Bellaousov, S., Reuter, J. S., Seetin, M. G., & Mathews, D. H. (2013). RNAstructure: web servers for RNA secondary structure prediction and analysis. Nucleic acids research, 41(W1), W471-W474.


2. Markham, N. R., & Zuker, M. (2008). UNAFold. In Bioinformatics (pp. 3-31). Humana Press.


3. J. N. Zadeh, C. D. Steenberg, J. S. Bois, B. R. Wolfe, M. B. Pierce, A. R. Khan, R. M. Dirks, N. A. Pierce. NUPACK: analysis and design of nucleic acid systems. J Comput Chem, 32:170–173, 2011. (pdf)


4. Skouridou, V., Jauset-Rubio, M., Ballester, P., Bashammakh, A. S., El-Shahawi, M. S., Alyoubi, A. O., & O’Sullivan, C. K. (2017). Selection and characterization of DNA aptamers against the steroid testosterone. Microchimica Acta, 184(6), 1631-1639.


5. Contreras Jiménez, G., Eissa, S., Ng, A., Alhadrami, H., Zourob, M., & Siaj, M. (2015). Aptamer-based label-free impedimetric biosensor for detection of progesterone. Analytical chemistry, 87(2), 1075-1082.


6. Neves, M. A., Slavkovic, S., Churcher, Z. R., & Johnson, P. E. (2017). Salt-mediated two-site ligand binding by the cocaine-binding aptamer. Nucleic acids research, 45(3), 1041-1048.


7. Liu, Z., Luo, D., Ren, F., Ran, F., Chen, W., Zhang, B., ... & Chen, Q. (2019). Ultrasensitive fluorescent aptasensor for CRP detection based on the RNase H assisted DNA recycling signal amplification strategy. RSC advances, 9(21), 11960-11967.


8. Denman, R. B. (1993). Using RNAFOLD to predict the activity of small catalytic RNAs. Biotechniques, 15(6), 1090-1095.


9. Bernhart, S. H., Tafer, H., Mückstein, U., Flamm, C., Stadler, P. F., & Hofacker, I. L. (2006). Partition function and base pairing probabilities of RNA heterodimers. Algorithms for Molecular Biology, 1(1), 1-10.


10. Ketterer, S., Fuchs, D., Weber, W., & Meier, M. (2015). Systematic reconstruction of binding and stability landscapes of the fluorogenic aptamer spinach. Nucleic acids research, 43(19), 9564–9572. https://doi.org/10.1093/nar/gkv944


11. Yildirim, I., Park, H., Disney, M. D., & Schatz, G. C. (2013). A dynamic structural model of expanded RNA CAG repeats: a refined X-ray structure and computational investigations using molecular dynamics and umbrella sampling simulations. Journal of the American Chemical Society, 135(9), 3528-3538.


12. Autour, A., Westhof, E., & Ryckelynck, M. (2016). iSpinach: a fluorogenic RNA aptamer optimized for in vitro applications. Nucleic acids research, 44(6), 2491–2500. https://doi.org/10.1093/nar/gkw083


13. http://www.mdtutorials.com/gmx/umbrella/07_analysis.html


14. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathan, S. A., & Karplus, M. (1983). CHARMM: a program for macromolecular energy, minimization, and dynamics calculations. Journal of computational chemistry, 4(2), 187-217.


15. Warner, K. D., Chen, M. C., Song, W., Strack, R. L., Thorn, A., Jaffrey, S. R., & Ferré-D'Amaré, A. R. (2014). Structural basis for activity of highly efficient RNA mimics of green fluorescent protein. Nature structural & molecular biology, 21(8), 658–663. https://doi.org/10.1038/nsmb.2865


16. Colasanti, A. V., Lu, X. J., & Olson, W. K. (2013). Analyzing and building nucleic acid structures with 3DNA. JoVE (Journal of Visualized Experiments), (74), e4401.


17. http://parts.igem.org/Part:BBa_K1330000


18. Wee, E. J., & Trau, M. (2016). Simple isothermal strategy for multiplexed, rapid, sensitive, and accurate miRNA detection. Acs Sensors, 1(6), 670-675.


19. Clint Moody, Heather Newell, Hendrik Viljoen, A mathematical model of recombinase polymerase amplification under continuously stirred conditions, Biochemical Engineering Journal, Volume 112, 2016, Pages 193-201, ISSN 1369-703X, https://doi.org/10.1016/j.bej.2016.04.017.


20. Stögbauer T, Windhager L, Zimmer R, Rädler JO. Experiment and mathematical modeling of gene expression dynamics in a cell-free system. Integrative Biology. 2012 May 1;4(5):494-501.


21. Morin JA, Cao FJ, Lázaro JM, Arias-Gonzalez JR, Valpuesta JM, Carrascosa JL, Salas M, Ibarra B. Mechano-chemical kinetics of DNA replication: identification of the translocation step of a replicative DNA polymerase. Nucleic acids research. 2015 Apr 20;43(7):3643-52.