Difference between revisions of "Team:Austin LASA/Model"

Revision as of 08:53, 17 October 2018

const i = f => h(g.MathJax.Node, {inline: true, formula: f});
const r = i => h('a', {href: '#ref_' + i}, i);
h(g.Page, {title: 'Model', prev: 'https://2018.igem.org/Team:Austin_LASA/Human_Practices', next: 'https://2018.igem.org/Team:Austin_LASA/Applied_Design', selector: [5, 1]},
  h('p', null, 'The LASA iGEM Team carried out the following modeling to count towards to Gold Medal Model criterion and Best Model Award:'),
  h('br'),
  h(g.Section, {title: 'LAMP Modelling'},
    h('p', null, 'In order for our Cas12a-based assay to be used as an effective point-of-care diagnostic for HIV, we need a portable and inexpensive means of amplifying viral DNA. We chose to work with LAMP (Loop-Mediated Isothermal Amplification) because of its high selectivity, rapid amplification, and isothermal nature (rendering a thermal cycler unnecessary).'),
    h('p', null, 'After qualitatively narrowing down our pool of potential LAMP primers from four to six (See the graphs on our Design, Development, and Results page), we needed to choose which of the remaining four was the most effective in amplifying our HIV sample. We used kinetic modelling to compare the efficiency of each set of primers with the purified Bst enzyme. A recent article by  Subramanian and Gomez proposed an empirical kinetic model for LAMP based on a generalized logistic curve [', r(1), ']. No ab initio model has been developed for the complex LAMP mechanism, but the logistic curve approximation can still be understood mechanistically in terms of the “competition between the so-called extended cauliflower-like structures and the complementary dumb bell structures in the cycling amplification step” [', r(1), '] during the LAMP reaction.'),
    h(g.MathJax.Provider, null,
      h('p', null, 'The model proposed by Subramanian and Gomez is of the form:'),
      h(g.MathJax.Node, {formula: 'y(t) = a + \\frac{(k−a)}{(1+e^{−b(t−m)})},'}),
      h('p', null, 'where ', i('y(t)'), ' is the concentration of the amplicon at ', i('t'), ', ', i('a'), ' is the starting concentration,  is the maximum concentration, ', i('m'), ' is the time at which maximum growth occurs, and ', i('b'), ' is a free parameter representing how steep the growth is. We fit our data to this model using SciPy’s curve_fit function. It is also worth noting that our data and fitted parameters are actually in units of fluorescence, not concentration. We assumed that the two were proportional and worked in terms of fluorescence instead because that was the data we had readily available.'),
      h('p', null, 'Once the model parameters have been obtained, we can compute ', i('T_p'), ' by ', i('T_p = m-\\frac{2}{b}'), ' [', r(1), '].'),
      h('p', null, 'Subramanian and Gomez’s model also includes an unambiguous quantitative means of calculating the time to positive (', i('T_p'), '), which is analogous to threshold cycling time for PCR. We can then establish a relationship between ', i('T_p'), ' and initial concentration of sample DNA, which is expected to be linear in accordance with Subramanian and Gomez’s data. For each of our four chosen primer sets (14, 573, 11, 12), we obtained ', i('T_p'), ' at the initial DNA concentrations of 10, 100, and 1000 fg/μL. We obtained the following data:'),
      h(g.Image, {src: 'https://static.igem.org/mediawiki/2018/f/f9/T--Austin_LASA--Tp_8_2.svg', position: 'center'}),
      h(g.Image, {src: 'https://static.igem.org/mediawiki/2018/0/02/T--Austin_LASA--Tp_8_4.svg', position: 'center'}),  
      h(g.Image, {src: 'https://static.igem.org/mediawiki/2018/6/6b/T--Austin_LASA--Tp_8_6.svg', position: 'center'}),  
      h(g.Image, {src: 'https://static.igem.org/mediawiki/2018/0/06/T--Austin_LASA--Tp_8_8.svg', position: 'center'}),    
      h('p', null, 'The data for primer sets 14, 573, and 12 were clearly linear as expected, and the linear regression for primer set 11 is still close enough to give a meaningful value for slope. We use the slopes of these regressions as a measure of the effectiveness of the primer set, with a higher magnitude corresponding to a greater primer effectiveness. Primer set 14 had a significantly higher slope magnitude than the other sets, so we can conclude that it will be the most effective primer for amplifying our HIV sample with LAMP.')
    )
  ),
  h(g.Section, {title: 'Future Work'},
    h('p', null, 'In the future, we would like to analyze the kinetics of Cas12a reactions carried out directly on LAMP amplicons. This would allow us to verify the sensitivity for the combined assay predicted by our model. Such a combined reaction would also be novel to our knowledge, and we’re interested in examining how the placement of the target sequence on the LAMP amplicons would affect the Cas12a kinetics. Although we’ve already collected data on both target sequences, we haven’t collected this data for a Cas12a reaction run on LAMP amplicons, and only the LAMP amplicons have the differing linear and loop segments, so we were unable to look into this with our current data.'),
    h('p', null, 'We were also interested exploring the kinetics of LAMP carried out with cellular reagents but didn’t have time to collect sufficient data for analysis. If we obtain this data in the future, we can apply essentially the same analysis as above to determine relative primer efficiency with cellular reagents. We could also compare how accurately Subramanian and Gomez’s model fits cellular reagent amplification curves and how well if fits purified enzyme data, which could provide us insight into how much background noise is introduced by the cellular reagents.')
  ),
  h(g.Section, {title: 'References'},
    h('p', null, '[', h('a', {id: 'ref_1'}, '1'), '] Subramanian S, Gomez RD (2014) An Empirical Approach for Quantifying Loop-Mediated Isothermal Amplification (LAMP) Using Escherichia coli as a Model System. PLoS ONE 9(6): e100596. ', h('a', {href: 'https://doi.org/10.1371/journal.pone.0100596'}, 'https://doi.org/10.1371/journal.pone.0100596'))
  )
);