DISCUSSION AND CONCLUSION
RNA Isolation and Purification:
Although the characterized mid-log phase optical density reading did not give as large of a concentration of yeast in the broth as a late-log phase growing period would, it could be said that the initial concentration values for both mutant and wild type yeast were adequate (Table 1). From the initial concentration of yeast, the target objective was to isolate as much pure RNA as possible. Because a typical yeast cell contains 1.2pg of RNA, and detrimental RNases can be found in almost every nook of the laboratory, it was extremely important to take care and be precise in the initial steps of the microarray experiment. As Table 1. in the results section indicates, the concentrations of isolated RNA were observed to be inadequate. The concentration values read on the spectrophotometer had two implications. First, when running the gel to check for RNase degradation and DNA impurity, a very small amount of thought to be sample, approximately 3ug opposed to the optimal 10ug, could be loaded into the wells before overflowing became an issue. This may have resulted in lower intensity reading on the gel, which could have caused an inaccurate description of what indeed was isolated (Figure 2.). Although this was a possibility by observing the agarose gel and the banding sequence pattern, it was in fact not the case. Instead, the gel electrophoresis gave good clues about the state of the RNA. If one is to examine Figure 2. in the results section, they can see three things about the gel. The Ribosomal RNA 28s 18s banding pattern is seen to be intact, which along with the lack of smearing indicated that the RNA kit, along with proper sanitation techniques, worked in removing proteins that could potentially degrade RNA. This result was promising, but there was a problem with the gel. At the bottom of both lanes, a bright entity was seen, which most likely represents DNA impurity. Because of the skew in the A260/A280 purity ratios, most likely due to the kit as noted by the irregularity in previous and current colleagues ratios, this gel was the only insight to what the isolated RNA sample was in fact containing. From the analysis of the RNA isolation techniques, it can be concluded that the RNA was intact, and was able to be used for the reverse transcription of messenger RNA; however, purity was lacking. This lack in purity and the skew it caused in loading of what was thought to be correct amounts of RNA may have weakened the intensities of the microarray causing inaccuracies.
In order to prepare the microarray, reverse transcription was used to make a cDNA library for the two different yeast samples. After the cDNA was prepared, the microarray slide was ready to be hybridized by the library. Two procedural errors occurred during the steps of microarray preparation. First, during application of the photo sensitive dies to the cDNA bound microarray, the initial attempt was unsuccessful due to a switch in dye vials with a DNA primer. Because of this mistake, lab techs had to come in late at night to right the wrong. Since dye administration was performed to multiple slides at the same time, possible photo bleaching and human error could have occurred. This may have resulted in a skew of the pixel count. During the final washing of the microarray, the beginning washes were performed with a more concentrated form of SSC solution. The problems in using the stronger concentration may have resulted in the washing off of the hybridized cDNA and/or dyes which, again, could be the cause of the lower intensity in the slides. Despite the human errors in slide preparation, the microarray was prepared and sent to Davidson college for scanning.
Microarray Analysis
Upon receiving the slides from Davidson college, it was noted that the green dyes in both the green ZMS1Δ mutant microarray and green wild-type dye reversal microarray had a much larger intensity of the green pigmented dye (Figure 3). This led the experiment to progress in a cross-comparison of the two green pigments. Using this initial cross-comparison, the global normalization scatter plot was generated with the green labeled mutant pixel numbers being on the x-axis and the green labeled wild-type on the y-axis (Figure 4). From this scatter plot, a correlation coefficient of less that 0.01 was calculated and the line of best fit was almost parallel to the x-axis. These results were indicative of the varying degree of concentrations between the two microarray slides and the uselessness of the graph generated. From this point, another global normalization was performed; however, this time the green ZMS1Δ microarray's pixels, both red and green, were utilized. From the scatter plot generated, a line of best fit much closer to the angle of 45o relative to the x-axis was produced, along with a better correlation coefficient of 0.7106 (Figure 5). The global normalization showed the definite skew of green pixels compared to red, which was due to photobleaching or a damaged red dye. Nonetheless, the green ZMS1Δ microarray normalization was much better than any other data produced. To eliminate any genes that had significant background noise, an empty spot control background check was performed. By taking the average pixel number of the empty spots and calculating two standard deviations up from that number, a cut off was obtained. This cut off filtered out any erroneous gene expression levels that were due to the difference in background pixel numbers. After performing these normalizations, genes that were thought to be induced or repressed were selected (Table 2).
From the filtration methods coming from normalization, expression levels from the green mutant could be determined from the ratio of green mutant pixels over red wild-type pixels. To easier see the amplification or repression of genes, the log base 2 of the ratios were determined. From these values, twenty six genes in the mutant yeast were identified as having higher levels of gene expression (Table 3). One gene in particular, YDR059C, showed an increase in expression levels and was involved in the oxidative stress pathway. As for the other 25 genes, there was an unknown reason for their increase in amplification since they were not involved in the prevention of oxidative stress or they had an unknown function. These unknown function genes, if data was better, would be an interesting pathway for analyzation. A cross-examination of the induced genes showed 13 genes, not including YDR059C, to have similar expression patterns. Again, these genes were not involved in the oxidative stress pathway. As far as repression in the gene expression levels, only 3 genes, all of unknown function, had a negative log base 2 value and did not match with the gene expression levels in the cross-examination.
Although there was a change in expression levels in the green ZMS1Δ mutant, there was too much randomized data and the conclusions were not substantial due to the human error as stated previously along with errors inherent in microarray analysis. If more experiments were conducted and result showed the fluctuation in expression levels of genes that were linked to the oxidative stress pathway then the data would be more viable. In conclusion, it can be said that there was a change in ZMS1Δ gene expression compared to that of wild-type, however it cannot be concluded that this change in expression level was due to the lack of ZMS1Δ transcription factor or microarray error.
Title Page Introduction Methods Results References
Contact Information: