Talk:20.109(S10):Data analysis (Day8)

T/R data

Below is your calcium titration data. Please fill in the second table by Saturday, if not today, so your classmates can compare their data to yours.

If your data is puzzling/noisy, note that both T/R Orange and T/R Purple have good WT and M124S data that you can look at.

Day 7 data file	Data taken after 24 hr settling period (optional)
Media:S10_M2D7_TR-Blue.txt	Media:S10_M2D7_TR-Blue-repeat.txt
Media:S10_M2D7_TR-Green.txt
Media:S10_M2D7_TR-Orange.txt
Media:S10_M2D7_TR-Pink.txt
Media:S10_M2D7_TR-Purple.txt
Media:S10_M2D7_TR-Red.txt	Media:S10_M2D7_TR-Red-repeat.txt
Media:S10_M2D7_TR-Yellow.txt	Media:S10_M2D7_TR-Yellow-repeat.txt

Please be mindful of significant figures when you fill in this table. Do not state far more precise numbers than you can believe.

Group Colour	X#Z Mutant	Most believable KD; n for WT	Most believable KD; n for M124S	Most believable KD; n for X#Z mutant	Comments
Pink	G59P	0.4974 , n= 8.52	0.7130, n= 3.19	0.5275, n= 15.68	We eliminated one of our WT series since we identifies significant noise due to pipetting error. We chose the model from Part 2 for our most believable values, since it takes into account CaM's cooperative behavior.
Purple	N137F	0.4534, n=9.8503	0.8724, n=3.8746	0.5312, n=21.2636	We chose the second model because the results most closely correlated to our predicted Excel data, and also the numbers were most believable and not completely outrageous (like model 3).
Blue	D129P	0.4431; 8.3	0.9535; 2.1	0.8092; 3.5	We chose the second model because the results correlated almost exactly with our predicted Excel data, factored in the cooperativity of CaM, and the absolute values of the residuals were closest to zero. The Hill coefficient shouldn't exceed 4..so we're looking into that.
Green	G23P	0.4606, 12.1895	0.3347, 0.8646	0.7678, 3.9190	For our M124S data, our average curve looked pretty awful, so we just took the normalized data of one of our data sets. We chose the part 2 model, which had Kd values close to those of the excel spreadsheet, accounted for cooperativity, and incorporated more data points.
Red	D20R	0.3224 ; 12	0.9284 ; 0.3	0.1853 ; 0	We adopted the second model in our analysis since it was the one most representative of the data. We got 12 for our n for WT, and 0 (actuallu -0.04) for our mutant. This comes in harmony with our observation that our protein remained fluorescent (i.e. did not bind to calcium) :)
Yellow	G23W/G25W	.5864; n=2.52576	.6862;n=2.3320	-.3936;n =2.0340	Our kD values for mutation 2 (our mutant) were nonsensical because the model used was not a good fit for our data. We received consistently high fluor. for all Ca concentrations. An alternative fit, using binding fraction = fluorescence (as opposed to using binding fraction =1 -fluorescence) gave us the following values: wt kd= .586; m124s kd =.686; g23w/g25w kd= .2878. These values are more reasonable, but they are not trustworthy because they are based on very few samples, and we had to get rid of an outlier to generate the alternative fit function.
Orange	G25P	KD=.4484 n=9.3825	KD=.9536 n=2.1191	KD=.5555 n=-.3638	We used model 2. It fit the control and M124S best. It was okay for G25P, but didn't perfectly match. However, there was little overall change in fluorescence and none of the other models fit well.

W/F data

Below is your calcium titration data. Please fill in the second table by Sunday, if not today, so your classmates can compare their data to yours.

If your data is puzzling/noisy, note that both T/R Orange and T/R Purple have good WT and M124S data that you can look at.

Note: Orange and Red groups, please use the bottom set (not the top set) of data in your file - something weird happened when it was saved.

Day 7 data file	Data taken after 24 hr settling period (optional)
Media:S10_M2D7_WF-Blue.txt
Media:S10_M2D7_WF-Green.txt
Media:S10_M2D7_WF-Orange.txt	Media:S10_M2D7_WF-Orange-repeat.txt
Media:S10_M2D7_WF-Pink.txt	Media:S10_M2D7_WF-Pink-repeat.txt
Media:S10_M2D7_WF-Purple.txt
Media:S10_M2D7_WF-Red.txt
Media:S10_M2D7_WF-Yellow.txt

Please be mindful of significant figures when you fill in this table. Do not state far more precise numbers than you can believe.

Group Colour	X#Z Mutant	Most believable KD; n for WT	Most believable KD; n for M124S	Most believable KD; n for X#Z mutant	Comments
Blue	Q104R	0.4490; 9.77	0.8001; 5.11	0.4451; 9.21	We chose the second model because it was the one that was the one that most correlated with our result. We had a little noise on the M124S data, but it was before the transitioin region of the data.
Yellow	E67K	0.1896; 1.7389	0.5274; 7.3622	0.4045; 4.7975	We used model 2 as the figures show the best fit with it. It also matched our predictions from our excel plots the best. Our wild type data appears inaccurate compared with T/R values, mainly because two of our points did not match the expected trend well at all.
Pink	E84G	.5071; 10.15	.7544; 2.54	.5589; 17.66	We used the second model because it gave a more accurate representation of our data. The residual plot of the 2nd model showed a more random plot than the 1st model.
Purple	D20R	0.4638; 6.8	0.8068; 4.1	0.4700; -Infinity	We used the second model because it gave a more accurate representation of our data. Both appeared sigmoidal and the M124S mutation appeared to decrease the affinity for calcium. Neither model fit our D20R mutant well, as it did not exhibit traditional cooperativity.
Red	E84K	0.44; 11.23	0.86; 2.30	0.55; 16.63	We used the model 2 data, since it have smaller residual values, and it best match out predicted data from excel, compared to model 1. Model 3 results are not too ideal because they yield too much error (little data sets in the transition region) We eliminated the first 4 values of 1 replicate of M124S because we believe there is pipetting error.
Orange	E84G	.4795; 15.4431	1.0119; .3923	.5583; 24.2170	We used the second model because it seemed to more accurately represent our data. We eliminated initital calcium concentrations for wild type and m124s since it was inconsistent with class data.
Green	D22H	0.428; 6.63	0; 0.02	1.41; 0.95	We used the second model because it had small abs(residuals). Model 3 results contradicted Model 2 results, but we trusted the Model 2 results better because they used more of the data. Our initial predictions and eyeball Excel estimations matched Model 2 results better. Note: our M124S data was extremely erratic and unexpected, so our Kd and n values are probably not reliable.