PubMed Health⌕ Search

Biomedical subjects

Michael C O'Neill

Publications and source records attributed to Michael C O'Neill.

3 recordsLinked to original sources
PubMed2003

Neural network analysis of lymphoma microarray data: prognosis and diagnosis near-perfect.

BACKGROUND: Microarray chips are being rapidly deployed as a major tool in genomic research. To date most of the analysis of the enormous amount of information provided on these chips has relied on clustering techniques and other standard statistical procedures. These methods, particularly with regard to cancer patient prognosis, have generally been inadequate in providing the reduced gene subsets required for perfect classification. RESULTS: Networks trained on microarray data from DLBCL lymphoma patients have, for the first time, been able to predict the long-term survival of individual patients with 100% accuracy. Other networks were able to distinguish DLBCL lymphoma donors from other donors, including donors with other lymphomas, with 99% accuracy. Differentiating the trained network can narrow the gene profile to less than three dozen genes for each classification. CONCLUSIONS: Here we show that artificial neural networks are a superior tool for digesting microarray data both with regard to making distinctions based on the data and with regard to providing very specific reference as to which genes were most important in making the correct distinction in each case.

Algorithms↗
PubMed2002

Evidence for "pre-recruitment" as a new mechanism of transcription activation in Escherichia coli: the large excess of SoxS binding sites per cell relative to the number of SoxS molecules per cell.

In response to the oxidative stress imposed by redox-cycling compounds like paraquat, Escherichia coli induces the synthesis of SoxS, which then activates the transcription of approximately 100 genes. The DNA binding site for SoxS-dependent transcription activation, the "soxbox," is highly degenerate, suggesting that the genome contains a large number of SoxS binding sites. To estimate the number of soxboxes in the cell, we searched the E. coli genome for SoxS binding sites using as query sequence the previously determined optimal SoxS binding sequence. We found approximately 12,500 sequences that match the optimal binding sequence under the conditions of our search; this agrees with our previous estimate, based on information theory, that a random sequence the size of the E. coli genome contains approximately 13,000 soxboxes. Thus, fast-growing cells with 4-6 genomes per cell have approximately 65,000 soxboxes. This large number of potential SoxS binding sites per cell raises the interesting question of how SoxS distinguishes between the functional soxboxes located within the promoters of target genes and the plethora of equivalent but nonfunctional binding sites scattered throughout the chromosome. To address this question, we treated cells with paraquat and used Western blot analysis to determine the kinetics of SoxS accumulation per cell; we also determined the kinetics of SoxS-activated gene expression. The abundance of SoxS reached a maximum of 2,500 molecules per cell 20 min after induction and gradually declined to approximately 500 molecules per cell over the next 1.5 h. Given that activation of target gene expression began almost immediately and given the large disparity between the number of SoxS molecules per cell, 2,500, and the number of SoxS binding sites per cell, 65,000, we infer that SoxS is not likely to activate transcription by the usual "recruitment" pathway, as this mechanism would require a number of SoxS molecules similar to the number of soxboxes. Instead, we propose that SoxS first interacts in solution with RNA polymerase and then the binary complex scans the chromosome for promoters that contain a soxbox properly positioned and oriented for transcription activation. We name this new pathway "pre-recruitment."

Bacterial Proteins↗