Microarray is a powerful technique that allows researchers to analyze tens of thousands of genes in a relatively short period of time. This recently developed technique enables us to quickly enhance our knowledge about gene expression patterns in patients as well as in laboratory samples. When we started using this technique about six years ago it was still in its infancy. Several drawbacks were observed that needed to be addressed before our microarray data could be confidently published. Because the problems described in our paper are worth consideration before publishing microarray data, several authors have acknowledged the importance of our paper and have cited it in their own manuscripts.

Our paper describes some potential problems associated with a newly developed technique, microarray. Several investigators are publishing their data without validating their results by other methods such as real-time PCR, Northern Blot analysis, and RNase protection assays. In addition, we observed that there is no uniform platform for microarray technology. Different types of microarrays give rise to different results. We believe that the problems described in our paper help other researchers publish more meaningful microarray data.

Microarray is a powerful tool that can aid in analyzing the expression of tens of thousands of individual genes in biological samples. It also helps to identify differentially expressed genes and in the classification of diseases. There are different types of microarrays such as cDNA microarray (offered by several vendors) and oligonucleotide probe arrays (Affymetrix). There is no uniform platform for this technique. When we were studying the differential expression of genes in large granular lymphocyte leukemia by using different types of microarrays, we found several problems. Some of the probes spotted on the microarray were of completely different sequences than what was reported by the manufacturer. Some of the probes were not specific and some of them lacked specificity for different gene isoforms. Several discrepancies were noted when cDNA microarray and oligonucleotide probe arrays were compared. Our paper strongly suggests that microarray results must be confirmed with other methods like Northern blots, RT-PCR, or RNase protection assays. Based on our results, we suggest the introduction of more uniform systems and verification of data by other methods before publishing data.

We were interested in identifying differentially expressed genes in large granular lymphocyte (LGL) leukemia. Since microarray is a powerful methodology, we used the microarray technique to identify differentially expressed genes in LGL leukemia. We utilized two different types of arrays. One was the Incyte Genomics cDNA microarray and the other was Affymetrix oligonucleotide array. These techniques were just emerging at the time; therefore we decided to verify our results as well as the sequences of the cDNA that were spotted on the array. We found several discrepancies. We also compared the results obtained by two different methods and found some disagreement. All the problems we encountered in our microarray analysis were recorded and published.