# I have done a SDS-PAGE to find the mass of the unknown protein... - (Dec/05/2006 )

I have created a graph to find the mass the of the unknown protein (couldnt attach the graph ).

I have to comment on the shape of the calibration curve and the precision with which I can determine the protein molecular masses. The problem is I dont understand!! What do I comment?!!

Pleeease pleeease pleeeeeeeease help!

-muskaan-

QUOTE (muskaan @ Dec 5 2006, 05:26 PM)
I have created a graph to find the mass the of the unknown protein (couldnt attach the graph ).

I have to comment on the shape of the calibration curve and the precision with which I can determine the protein molecular masses. The problem is I dont understand!! What do I comment?!!

Pleeease pleeease pleeeeeeeease help!

the standard plo is still a classical Neville plot (launched in 1971 if I´m right)

-The Bearer-

you must use a logarithmic scale for the molecular weight, and then you should obtain a linear curve.

-Missele-

QUOTE (Missele @ Dec 5 2006, 06:55 PM)
you must use a logarithmic scale for the molecular weight, and then you should obtain a linear curve.

that is the Neville plot (log Mw/Rf)

-The Bearer-

The shape of the graph: useing a mathematical modelling program you'll find which equation "fits" your data. Linear? Exponential? You may need to transform your data (e.g. semi log, log-log). Once you know what equation fits (y=mx+c?) you can make assumptions and predictions. You can also see the range of linearity e.g. the concentration range that your technique gives adequate precision (see below).

As far as the precision is concerned, you need to run repeats at each concentration used in the standard curve and your protein. Go for triplicates at first. You calculate the standard deviation (SD) and the mean value then use those two to find the %Co-efficient of Variation (CV%).

The sd is sq root of ( (the mean squared) minus (the mean squared divided by n) all divided by n-1). The CV% is then sd/mean.
CV is the precision you quote. This in fact is an expression of total error so includes an assessment of your pipetting skills as well as random error in the technique. You should get very tight CVs for this (under 1%, maybe under 5%).

Another fairly useful technique is to plot precision curves. In this you measure the CVs of your protein at different concentrations (it'll be the same size but yhe spread of the triplicates may vary). Plot the CVs against the concentrations. You should get a lop-sided smiley graph! (Its harder to measure very small quantities accurately so CVs are high. At the high end, the method pops out of linearity).

Have fun.

-paraboxa-