spacer image
spacer image

Expression Profiling Quality Control Measures
spacer image

To ensure that all the profiles in our public access database are comparable (see navigate repository), we have spent considerable time and effort identifying sources of variability in microarray data (Bakay et al. 2002a), and developing quality control parameters (QC) and standard operating procedures (SOP). SOPs developed within the center employ the QC checkpoints at various stages in the typical microarray workflow from total RNA extraction to GeneChip® hybridization and scanning. This standardized methodology for quality assessment has been shown to result in the generation of reliable and reproducible data. A summary of the threshold values for each QC criteria follows, with a more detailed explanation of each of the measures in the next section.

Summary of QC criteria:

  1. RNA: clearly visible 18S and 28S RNA bands, and A260/280 ratio of >1.8.
  2. Input total RNA:
    1. Standard (one round): 5-10 micrograms for one round amplifications
    2. Two-round: 100-200 ng of total RNA input into first round, 100 ng converted cRNA/cDNA into second round.
  3. IVTs:
    1. Standard (one round): ouput cRNA = 4-10 fold adjusted amplification
    2. Two round: Round 1: 4-10 fold adjusted amplification. Round 2: 400-fold adjusted amplification.
  4. Biotinylated cRNA integrity: 500-3,000 bp distribution
  5. Biotinylated cRNA fragmentation: 50-200 bp distribution
  6. Hybridization (chip) controls:
    1. Visual inspection of image shows no chip defects
    2. Percent present calls >30% (A chips)
    3. Background intensity: <150 (PMT = 1,800)
    4. Scaling Factor: <5.0 (PMT = 1,800)
    5. Spike in controls: “present”, BioB<BioC<BioD<CreX
  7. cDNA integrity: GAPDH 3’/5’ ratio <3.0

The following descriptions summarize our experience with the above parameters.

  1. RNA integrity and purity: Extracted total RNA is evaluated on a 1% agarose gel or Agilent Bioanalyzer for presence of 28S and 18S rRNA bands. Ideally, the intensity of the 28S band should be twice the intensity of the 18S. RNA which is partially degraded will appear smeared, and lack rRNA bands. RNA concentration is quantitated by UV spectrophotometry at an absorbance of 260, or using an Agilent Bioanalyzer. The A260/280 ratio is used as a measure of RNA purity. RNA with an A260/280 ratio of 1.8-2.1 is considered acceptable for expression profiling. RNA that does not meet these QC criteria typically fail the IVT fold-amplification threshold at a high rate (see below).
  2. RNA quantity: For single round amplifications (one IVT reaction), between 3 and 10 micrograms of total RNA is used (see required fold-amplifications below). For two round IVTs from small samples, 100-200 ng of total RNA is input into the first round, and 100 ng of adjusted cRNA input into the second round. Note that a common reason for failure of two round amplifications is the use of too much input cRNA into the second round.
  3. IVT fold-amplification: We have found that the single most critical QC measure is the efficiency of amplification from the in vitro transcription reaction. Inadequate amplification reactions lead to low amounts of biotinylated cRNA, and these samples typically fail subsequent chip-based QC measures. Inefficient IVT reactions are also the major cause for rejection of samples, as they can be variable. We required that each IVT (from either 1 round or 2 round amplifications) give a minimum of a 400-fold amplification of polyA+ mRNA. For 1 round amplifications, each sample must achieve four-fold amplification from the starting total RNA amount to the adjusted cRNA amount (as polyA+ is only ~1% of total RNA, this translates to a 400-fold amplification of polyA+ mRNA). For example, a sample started from 6ug of total RNA must have no less than 30ug after the IVT (30ug of cRNA- 6ug starting material = 24ug LABELED RNA, i.e. four-fold amplification). Additionally, the Microarray Center tracks the IVT fold-amplification for all samples within a project for consistency. Samples with abnormally high or low amplifications are flagged in subsequent data analysis. For 2 round amplifications, 100-200 ng of input total RNA should give a minimum of 400 ng of unlabeled cRNA. When this is converted to double-stranded cDNA, 100 ng of this cDNA is input into the second biotinylated IVT, and must provide 40 micrograms of biotinylated cRNA. Note that the adjusted output of the first round is considered “pure polyA+ cDNA”, and thus there is a 400-fold amplification required.
  4. cRNA integrity: cRNA is evaluated for quality on an agarose gel or Agilent Bioanalyzer similar to the way that total RNA was assessed. The cRNA gel should show a smear of RNA ranging from 500-3000 bases. A large abundance of RNA around 100bp indicates degradation of the RNA while intensely bright wells indicate significant genomic DNA contamination.
  5. cRNA fragmentation: Proper base hydrolysis of the cRNA is also checked by agarose gel or Agilent Bioanalyzer. Fragments must range from 50-200bp in size. Smaller fragments indicate over fragmentation and will result in unreliable hybridization, larger fragments indicate insufficient fragmentation.
  6. Hybridization efficiency:
    7. GeneChip® hybridization is assessed by several parameters, namely: % present calls, background intensity, scaling factor and spike-in controls.
    1. Visual inspection of array hybridization: Microarray images are manually inspected for any chip defects (scratches, holes). Chips with obvious defects are discarded, and the hybridization solution run on a new chip.
    2. Percent present calls: As per Affymetrix suggestion, the Microarray Center requires that all human, murine and rat A chips have a minimum of 30% of the represented genes called “present”. B series chips are required to have a minimum of 20% present calls. Note that older U95 series arrays (B, C, D, E) have less present calls due to ambiguous ESTs and poorly performing probe sets (see Bakay et al. 2002b )
    3. Background intensity: On GeneChip® scanners with PMT values set at 1800, the Microarray center requires that all chips have a background intensity of no more than 150. On GeneChip® scanners with PMT values set at 18,000 the Microarray Center requires that all arrays have a background intensity of no more than 1500. Additionally, the Microarray Center tracks the background intensities of all GeneChips® within a project and flags samples with background intensities outside the range of others in the project.
    4. Scaling factor: Also as per Affymetrix suggestion, the Microarray Center requires that all GeneChips® used for analysis have a scaling factor of no more than 5.0 when GeneChip® scanners are tuned at high (18000) or low (1800) PMT voltages with a target intensity respectively of 800 and 150.
    5. Spike-in controls: As described in the Affymetrix protocol, all GeneChip® hybridizations include the use of spike-in controls, namely, BioB, BioC, BioD, and CreX. The Microarray Center requires that all chips used in analysis have ‘present’ calls for BioC, BioD, and CreX. Additionally, hybridization efficiency is only considered acceptable when these genes are detected in increasing quantities: BioB<BioC<BioD<CreX.
  7. cDNA integrity: The quality of hybridized material is assessed by use 3’ and 5’ ends endogenously expressed housekeeping controls. Specifically the center routinely uses GAPDH for human, murine, and rat chips. Only GeneChips® with ratios <3.0 are considered analysis.
 
spacer image
goto topspacer image
spacer image
Grey dot spacer image