^ E& E3 Lp ¡¿j

no bgr subtr no bgr subtr lowest on average on mode of progenesis ADJUSTED boundary boundary non-spot (see above)

FIGURE 8.5. The top panel shows the variance in spot volumes across five copies of the same gel image following analysis with SameSpots, using the Progenesis background subtraction algorithm. Each bar represents the CV of one protein spot across the five identical replicates, and the bars are ordered according to increasing average spot volume. The bottom panel shows a box plot representing the CVs for the same experiment for each of the available background subtraction algorithms (lowest-on-boundary, average-on-boundary, mode-of-non-spot, and progenesis). In addition, the results from complete exclusion of background subtraction (no bgr subtr), as well as an adjustment of these results to correct for the decrease in CV caused by the larger spot volumes resulting from exclusion of background subtraction, are shown. Each box contains the interquartile range with the mean marked inside, and the whiskers show the range of CVs following exclusion of outliers.

the no-background-subtraction results was performed through dividing the spot volumes with the ratio of the no-background-subtraction and lowest-on-boundary spot volumes (Figure 8.5). The trend remains the same also after transformation, emphasizing the importance of evaluating the choice of background subtraction method because the quantity and nature of the variance in a system has direct implications on the statistical power. Considering that the biological variance within the control group is 40-50% [30], an additional technical variance of 10-20% caused by the analysis software will have a significant effect on the number of replicates needed. Based on a statistical confidence of 80% power and a 0.05 p-value, an overall variance of 50% would require n = 3-4, while a variance of 70% would require n = 8 [30].

The same analysis was also performed for the authentic 2D gel replicate set (i.e., including technical variance). The software-associated variance accounted for 12-16% of the total variance for all background subtraction methods offered in the SameSpots software—except mode-of-non-spot, which caused for 35% of the total variance. These results correspond well to the previous studies performed with Expression (PG200) [17]. In similar studies performed on DIGE gel images, the DeCyder software appeared to introduce 30% of the total variance (reported as the unexplained variance) [31], which corresponds well both with previous results from PDQuest [17] as well as the results presented here using the mode-of-non-spot algorithm.

Was this article helpful?

0 0
How To Win Your War Against Allergies

How To Win Your War Against Allergies

Not Able To Lead A Happy Life Because Of Excessive Allergies? Want To Badly Get Rid Of Your Allergy Problems, But Are Super Confused And Not Sure Where To Even Start? Don't Worry, Help Is Just Around The Corner Revealed The All-In-One Power Packed Manual Containing Ample Strategies And Little-Known Tips To Get Rid Of Any Allergy Problems That Are Ruining Your Life Learn How You Can Eliminate Allergies Completely Reclaim Your Life Once Again

Get My Free Ebook

Post a comment