Deterministic projection by growing cell structure networks for visualization of high-dimensionality datasets

Recent advances in clinical proteomics data acquisition have led to the generation of datasets of high complexity and dimensionality. We present here a visualization method for high-dimensionality datasets that makes use of neuronal vectors of a trained growing cell structure (GCS) network for the projection of data points onto two dimensions. The use of a GCS network enables the generation of the projection matrix deterministically rather than randomly as in random projection. Three datasets were used to benchmark the performance and to demonstrate the use of this deterministic projection approach in real-life scientific applications. Comparisons are made to an existing self-organizing map projection method and random projection. The results suggest that deterministic projection outperforms existing methods and is suitable for the visualization of datasets of very high dimensionality.     

Identification of nocardia species by restriction endonuclease analysis of an amplified portion of the 16S rRNA gene

Identification of clinical isolates of Nocardia to the species level is important for defining the spectrum of disease produced by each species and for predicting antimicrobial susceptibility. We evaluated the usefulness of PCR amplification of a portion of the Nocardia 16S rRNA gene and subsequent restriction endonuclease analysis (REA) for species identification. Unique restriction fragment length polymorphism (RFLP) patterns were found for Nocardia sp. type strains (except for the N. asteroides type strain) and representative isolates of the drug pattern types of Nocardia asteroides (except for N. asteroides drug pattern type IV, which gave inconsistent amplification). A variant RFLP pattern for Nocardia nova was also observed. Twenty-eight clinical isolates were evaluated both by traditional biochemical identification and by amplification and REA of portions of the 16S rRNA gene and the 65-kDa heat shock protein (HSP) gene. There was complete agreement among the three methods on identification of 24 of these isolates. One isolate gave a 16S rRNA RFLP pattern consistent with the biochemical identification but was not identifiable by its HSP gene RFLP patterns. Three isolates gave 16S rRNA RFLP patterns which were inconsistent with the identification obtained by both biochemical tests and HSP gene RFLP; sequence analysis suggested that two of these isolates may belong to undefined species. The PCR and REA technique described appears useful both for the identification of clinical isolates of Nocardia and for the detection of new or unusual species.     

The effects of sleep posture and sleep stage on apnea frequency

The Apnea Plus Hypopnea Index (A + HI) of 60 male positional sleep apneics was analyzed by sleep stage to determine if positional differences are limited to NREM sleep. Differences in apnea severity by sleep position were found to persist in REM sleep and to be of equal extent to those differences found in NREM sleep, despite the fact that there is also a significant increase in the frequency of apneic events associated with REM sleep. The positional effect persists in REM sleep, making treatments to control sleep posture a viable option.     

Percutaneous endoscopic trigonoplasty in children: long-term outcomes and modifications in technique

PURPOSE: Long-term outcomes of a minimally invasive method of correcting vesicoureteral reflux are presented with a discussion of the modification in our original technique. PATIENTS AND METHODS: A total of 29 children (46 refluxing ureters), 14 months to 18 years old, underwent percutaneous endoscopic trigonoplasty (PET) between December 1994 and June 1996. Follow-up ranged from 19 to 37 months. Reflux was grade 1 in 2, grade 2 in 16, grade 3 in 19, grade 4 in 8, and grade 5 in 1. The technique was a Gil-Vernet method in the first 23 patients and Cohen reimplantation in the last 6 patients. RESULTS: Resolution of reflux was observed to decrease from 63% to 47% with long-term (30-37 months') follow-up using the Gil-Vernet technique. Resolution was greater with unilateral reflux than bilateral reflux (83% v. 27%, respectively). The Cohen technique resulted in resolution of reflux in 83%; however, the operating time nearly doubled when compared with the Gil-Vernet group. In both groups, failure was unrelated to grade of reflux, age, operative sequence, or bladder instability. CONCLUSIONS: Although showing an improvement in resolution of reflux over the Gil-Vernet PET procedure, the Cohen PET reimplant has a lower success rate than traditional open operative reimplants. The PET also requires more operating time and two operating surgeons. Despite some advantages in the promptness of recovery, we do not recommend PET by either technique at this time. Future modifications may make this approach more tenable.