Problems in determining the results

Several problems can develop during the postanalysis interpretation of these results.

• No genes BLAST to finding: Some microarray results may be given as accession numbers for ESTs. Thus, to follow up these results, it is necessary to take the nucleotide sequence for the EST and BLAST that sequence against GenBank or UniGene looking for a "full-length" gene. In translating these accession numbers, however, there may be no gene that matches strongly enough. This can happen because the full-length gene may not be known yet, or due to inaccuracies in the EST consensus sequence. Gene findings with this problem may not be able to be followed up.

• Many genes BLAST to finding: Sometimes, many full-length genes may be found for a single EST, meaning the EST is nonspecific. At the time of the microarray construction, the EST may have been thought to be specific for one gene, but with progress in sequencing, this may no longer be true. To follow up these findings, many clones may need to be used in biological experiments.

• EST is no longer unique: Often, microarrays may have been designed with a particular sequence thought to be unique for a single gene, EST, or EST cluster. With progress in sequencing, however, it may be later realized that what was thought to be a single EST or gene cluster actually represents two or more unique clusters. It is often unclear as to which of the new clusters best maps to the older set of probes. Unfortunately, revision control is often not comprehensive in these accession number databases, and it becomes very difficult to link new clusters with old designators.

• Genomic sequence BLASTs, but not genes: Sometimes, gene sequences from positive microarray findings can be found in stretches of genomic sequence using BLAST. Additional bioinformatics (and biological) work then needs to be done to determine the gene boundaries before biological validation can take place.

Was this article helpful?

0 0

Post a comment