Profiled support vector machines for antisense oligonucleotide efficacy prediction

Background  

This paper presents the use of Support Vector Machines (SVMs) for prediction and analysis of antisense oligonucleotide (AO) efficacy. The collected database comprises 315 AO molecules including 68 features each, inducing a problem well-suited to SVMs. The task of feature selection is crucial given the presence of noisy or redundant features, and the well-known problem of the curse of dimensionality. We propose a two-stage strategy to develop an optimal model: (1) feature selection using correlation analysis, mutual information, and SVM-based recursive feature elimination (SVM-RFE), and (2) AO prediction using standard and profiled SVM formulations. A profiled SVM gives different weights to different parts of the training data to focus the training on the most important regions.     

Effect of nitrate-nitrogen limitation on photosynthesis of the diatom Phaeodactylum tricornutum Bohlin (Bacillariophyceae)

Abstract Nitrate limited growth of the diatom Phaeodactylum tricornutum in chemostat cultures produced marked changes in biochemical composition and a six-fold reduction in the specific growth rate. This was associated with a reduction in the carbon and chlorophyll a specific light saturated rates, with little effect on light limited photosynthesis. Variations in specific growth rate were quantitatively related to carbon specific net photosynthesis and maximum chlorophyll a specific light saturated rates were positively correlated with cell nitrogen contents. The correlation between nitrogen content and photosynthesis for P. tricornutum and the differential effect of nitrogen supply on the light response curve of photosynthesis is qualitatively and quantitatively similar to published results for terrestrial vascular plants. There was little change in the photon (quantum) yield of photosynthesis which was not significantly different from 0.125mol O₂ mol photon^-1 the theoretical upper limit based on the Z scheme, even under severe nitrate deficiency. The capacity to maintain a high photon yield under nitrate limitation is discussed in relation to the nitrogen requirements of the stromal and membrane components of the photosynthetic apparatus.     

Riparian vegetated buffer strips in water-quality restoration and stream management

• 1 A review is presented of the literature on riparian vegetated buffer strips (VBS) for use in stream-water-quality restoration and limitations associated with their use are discussed. The results are also presented of recent investigations on the effectiveness of a forested and a grass vegetated buffer strip for reducing shallow subsurface inputs of nutrients from agriculture to a stream in central Illinois, U.S.A.
• 2 Because riparian zones link the stream with its terrestrial catchment, they can modify, incorporate, dilute, or concentrate substances before they enter a lotic system. In small to mid-size streams forested riparian zones can moderate temperatures, reduce sediment inputs, provide important sources of organic matter, and stabilize stream banks.
• 3 Several questions on the utility and efficiency of vegetated buffer strips for stream restoration still remain unanswered, including: what types (grass v forest) are most efficient; do they become nutrient saturated; are they only temporary sinks; how does species composition influence effectiveness; and, what is the optimal width of buffer to facilitate nutrient reduction under different conditions?
• 4 Water samples were collected (1989–90) from lysimeters located at three depths (60, 120, and > 120cm) in an upland area planted in conventional row crops (corn and soybean) and in three adjacent riparian buffer treatments, a 39m wide grass buffer. a 16 m wide mature forested buffer, and a buffer planted in row-crops to the stream bank. Concentrations of dissolved and total phosphorus and nitrate-N in each sample were determined following major precipitation events over a seventeen month period.
• 5 Both the forested and grass VBS reduced nitrate-N concentrations in shallow groundwater (up to 90% reduction). On an annual basis the forested VBS was more effective at reducing concentrations of nitrate-N than was the grass VBS, but was less efficient at retaining total and dissolved P.
• 6 During the dormant season, both grass and forested buffer strips released dissolved and total P to the groundwater. The VBS apparently acted as a nutrient sink for much of the year, but also released accumulated nutrients during the remaining portion of the year. Periodic harvesting of plant biomass may reduce the amount of P released during the dormant season.
• 7 VBSs are not as effective in agriculture areas with tile drained fields. Alternative restoration practices such as discharging drain tiles into wetlands constructed parallel to the stream channel may prove to be a more effective means of controlling non-point-source agricultural inputs of nutrients in such areas.

     

Natural rehabilitation of stream fish populations in an Illinois catchment

• 1 An extreme drought in 1988 dried first- and second-order streams in the Salt Fork basin (1190 km²) in east-central Illinois. This event provided a natural experiment in which the natural rehabilitation of fish populations could be measured in the whole catchment following resumption of stream flows.
• 2 Fish were sampled before, during, and after the drought (1987–1990) throughout the basin (eighty-eight samples in July-September) using methods of known efficiency. Analyses of covariance [using log (distance from source) as the covariate] indicated no significant differences (P>0.2) of biomass or species richness between pre-drought and post-drought samples from sites desiccated during 1988 or among pre-drought, drought, and post-drought years among samples from perennial Streams. Therefore, recovery occurred within 1 year, but there was no indication of increased biomass or species richness in permanent streams resulting from fish moving down or remaining downstream during the drought.
• 3 Fish biomass per unit length of stream segment was modelled as a power function of distance of the segment from the source of each stream to estimate changes in biomass of the whole catchment. Before and subsequent to the drought, 90.0t of fish occupied the total 729km of stream length in the basin, compared with 74.61 in the drought year. Although, the drought affected 80% of total stream length, fish biomass was reduced by only 17% in the drought year because only the lower-order streams were desiccated and their normal biomass density was lower than in perennial streams.
• 4 In conclusion, it is expected that no expenditure on stocking fish would be necessary in restoration projects on these lower-order, surface runoff streams, providing that they are connected to permanently flowing streams that contain a full complement of species.

     

Restoration of lowland streams: an introduction*

• 1 This paper introduces the Lowland Streams Restoration Workshop that was held in Lund, Sweden in August 1991.
• 2 Attenders at the Workshop participated in working groups which discussed and reported on the state of knowledge of stream restoration and identified critical areas of information need. Currently, most restoration efforts are emission-orientated (i.e. waste-water management), while the imitation of the geomorphology or of the riparian vegetation of a quasi-natural or natural reference channel receives less attention.
• 3 Successful stream restoration requires a multidisciplinary approach within a holistic system framework. Monitoring the outcome of past, existing and future steam-restoration projects is required for information on the feasibility of alternative techniques and approaches.
• 4 It was recommended that systems in pristine condition serve as a point of reference and not as a goal for most stream restoration projects. Restoration goals must be carefully defined so that everyone at every level understands the aim of the project. At the very least, all restoration programmes should consider geomorphic, hydrological, biological, aesthetic, and water quality aspects of the system.
• 5 Restoration programmes should aim to create a system with a stable channel, or a channel in dynamic equilibrium that supports a self-sustaining and functionally diverse community assemblage; it should not concentrate on one species or group, except at the local level. Preserving the terrestrial -aquatic interface by setting aside riparian land corridors is critical to all stages of restoration. Additional information on the temporal and regional variability in important system processes and functions is needed.

     

Size distribution analysis of recombinant adenovirus using disc centrifugation

Recombinant adenovirus is one of the primary vectors for human gene therapy. However, the aggregation of unstable virus has been a recurring problem during the production of purified virus for human therapeutics. To facilitate the development of a robust manufacturing process for recombinant adenovirus vectors, a convenient and reliable size distribution analytical assay is necessary and we demonstrate here that disc centrifuge sedimentation is applicable to this purpose. Using the disc centrifuge system and the line start method, the assay can provide particle size distribution of adenovirus samples within 30 min. The assay can detect virus concentrations down to 0.01% (w/v) or 3 × 10¹¹ particles per ml. The apparent hydrodynamic diameter of recombinant adenovirus was determined to be about 0.063 μm. Furthermore, the disc centrifuge analysis was able to detect adenovirus dimers, trimers, and tetramers, consistent with a rigid sphere approximation for adenovirus, as well as a large aggregate of 0.35 μm. The appearance of viral aggregates is confirmed by increased light scattering based on A₃₂₀/A₂₆₀ ratios. The technique could be useful for monitoring the kinetics of aggregation for adenovirus and other DNA and RNA viruses in the submicron region. Therefore, this novel assay provides a critical tool for purification development of viral vectors for meeting therapeutic and research needs. Received 18 September 1997/ Accepted in revised form 15 May 1998     

Proliferation of direct repeats near the Oenothera chloroplast DNA origin of replication

The spacer between the 16S and 23S rRNA genes of the chloroplast DNA has been implicated as an origin of replication in several species of plants. In the evening primrose, Oenothera, this site was found to vary greatly in size, with plastid genomes (plastomes) being readily distinguished. To determine whether plastome "strength" in transmission could be correlated with variation at oriB, the 16S rRNA-trnI spacer was sequenced from five plastomes. The size variation was found to be due to differential amplification (and deletion) of combinations of sequences belonging to seven families of direct repeats. From these comparisons, one short series of direct repeats and one region capable of forming a hairpin structure were identified as candidates for the factor that could be responsible for the differences between strong and weak plastome types. Ample sequence variation allowed phylogenetic inferences to be made about the relationships among the plastomes. Phylogenetic trees also could be constructed for most of the families of direct repeats. The amplifications and deletions of repeats that account for the size variation at oriB are proposed to have occurred through extensive replication slippage at this site.      

Internal eliminated sequences interrupting the Oxytricha 81 locus: allelic divergence, conservation, conversions, and possible transposon origins

Internal eliminated sequences (IESs) often interrupt ciliate genes in the silent germline nucleus but are exactly excised and eliminated from the developing somatic nucleus from which genes are then expressed. Some long IESs are transposons, supporting the hypothesis that short IESs are ancient transposon relics. In light of that hypothesis and to explore the evolutionary history of a collection of IESs, we have compared various alleles of a particular locus (the 81 locus) of the ciliated protozoa Oxytricha trifallax and O. fallax. Three short IESs that interrupt two genes of the locus are found in alleles from both species, and thus must be relatively ancient, consistent with the hypothesis that short IESs are transposon relics. In contrast, TBE1 transposon interruptions of the locus are allele-specific and probably the results of recent transpositions. These IESs (and the TBE1s) are precisely excised from the DNA of the developing somatic macronucleus. Each IES interrupts a highly conserved sequence. A few nucleotides at the ends of each IES are also conserved, suggesting that they interact critically with IES excision machinery. However, most IES nucleotide positions have evolved at high rates, showing little or no selective constraint for function. Nonetheless, the length of each IES has been maintained (+/- 3 bp). While one IES is approximately 33 bp long, three other IESs have very similar sizes, approximately 70 bp long. Two IESs are surrounded by direct repeats of the sequence TTCTT. No other sequence similarities were found between any of the four IESs. However, the ends of one IES do match the inverted terminal repeat consensus sequence of the "TA" IESs of Paramecium. Three O. trifallax alleles appear to have been recipients in recent conversion events that could have been provoked by double-strand breaks associated with IES ends subsequent to IES transposition. Our findings support the hypothesis that short IESs evolved from ancient transposons that have lost most of their sequences, except those necessary for precise excision during macronuclear development.