Drosophila melanogaster has been established as a model organism for investigating the developmental gene interactions. The spatio-temporal gene expression patterns of Drosophila melanogaster can be visualized by in situ hybridization and documented as digital images. Automated and efficient tools for analyzing these expression images will provide biological insights into the gene functions, interactions, and networks. To facilitate pattern recognition and comparison, many web-based resources have been created to conduct comparative analysis based on the body part keywords and the associated images. With the fast accumulation of images from high-throughput techniques, manual inspection of images will impose a serious impediment on the pace of biological discovery. It is thus imperative to design an automated system for efficient image annotation and comparison.
Results
We present a computational framework to perform anatomical keywords annotation for Drosophila gene expression images. The spatial sparse coding approach is used to represent local patches of images in comparison with the well-known bag-of-words (BoW) method. Three pooling functions including max pooling, average pooling and Sqrt (square root of mean squared statistics) pooling are employed to transform the sparse codes to image features. Based on the constructed features, we develop both an image-level scheme and a group-level scheme to tackle the key challenges in annotating Drosophila gene expression pattern images automatically. To deal with the imbalanced data distribution inherent in image annotation tasks, the undersampling method is applied together with majority vote. Results on Drosophila embryonic expression pattern images verify the efficacy of our approach.
Conclusion
In our experiment, the three pooling functions perform comparably well in feature dimension reduction. The undersampling with majority vote is shown to be effective in tackling the problem of imbalanced data. Moreover, combining sparse coding and image-level scheme leads to consistent performance improvement in keywords annotation.
“Stoichioproteomics” relates the elemental composition of proteins and proteomes to variation in the physiological and ecological environment. To help harness and explore the wealth of hypotheses made possible under this framework, we introduce GRASP (http://www.graspdb.net), a public bioinformatic knowledgebase containing information on the frequencies of 20 amino acids and atomic composition of their side chains. GRASP integrates comparative protein composition data with annotation data from multiple public databases. Currently, GRASP includes information on proteins of 12 sequenced Drosophila (fruit fly) proteomes, which will be expanded to include increasingly diverse organisms over time. In this paper we illustrate the potential of GRASP for testing stoichioproteomic hypotheses by conducting an exploratory investigation into the composition of 12 Drosophila proteomes, testing the prediction that protein atomic content is associated with species ecology and with protein expression levels.
Results
Elements varied predictably along multivariate axes. Species were broadly similar, with the D. willistoni proteome a clear outlier. As expected, individual protein atomic content within proteomes was influenced by protein function and amino acid biochemistry. Evolution in elemental composition across the phylogeny followed less predictable patterns, but was associated with broad ecological variation in diet. Using expression data available for D. melanogaster, we found evidence consistent with selection for efficient usage of elements within the proteome: as expected, nitrogen content was reduced in highly expressed proteins in most tissues, most strongly in the gut, where nutrients are assimilated, and least strongly in the germline.
Conclusions
The patterns identified here using GRASP provide a foundation on which to base future research into the evolution of atomic composition in Drosophila and other taxa.
Improvements in sequencing technology now allow easy acquisition of large datasets; however, analyzing these data for phylogenetics can be challenging. We have developed a novel method to rapidly obtain homologous genomic data for phylogenetics directly from next-generation sequencing reads without the use of a reference genome. This software, called SISRS, avoids the time consuming steps of de novo whole genome assembly, multiple genome alignment, and annotation.
Results
For simulations SISRS is able to identify large numbers of loci containing variable sites with phylogenetic signal. For genomic data from apes, SISRS identified thousands of variable sites, from which we produced an accurate phylogeny. Finally, we used SISRS to identify phylogenetic markers that we used to estimate the phylogeny of placental mammals. We recovered eight phylogenies that resolved the basal relationships among mammals using datasets with different levels of missing data. The three alternate resolutions of the basal relationships are consistent with the major hypotheses for the relationships among mammals, all of which have been supported previously by different molecular datasets.
Conclusions
SISRS has the potential to transform phylogenetic research. This method eliminates the need for expensive marker development in many studies by using whole genome shotgun sequence data directly. SISRS is open source and freely available at https://github.com/rachelss/SISRS/releases.
Background:
Many pharmaceutical drugs are known to be ineffective or have negative side effects in a substantial proportion of patients. Genomic advances are revealing that some non-synonymous single nucleotide variants (nsSNVs) may cause differences in drug efficacy and side effects. Therefore, it is desirable to evaluate nsSNVs of interest in their ability to modulate the drug response.
Results:
We found that the available data on the link between drug response and nsSNV is rather modest. There were only 31 distinct drug response-altering (DR-altering) and 43 distinct drug response-neutral (DR-neutral) nsSNVs in the whole Pharmacogenomics Knowledge Base (PharmGKB). However, even with this modest dataset, it was clear that existing bioinformatics tools have difficulties in correctly predicting the known DR-altering and DR-neutral nsSNVs. They exhibited an overall accuracy of less than 50%, which was not better than random diagnosis. We found that the underlying problem is the markedly different evolutionary properties between positions harboring nsSNVs linked to drug responses and those observed for inherited diseases. To solve this problem, we developed a new diagnosis method, Drug-EvoD, which was trained on the evolutionary properties of nsSNVs associated with drug responses in a sparse learning framework. Drug-EvoD achieves a TPR of 84% and a TNR of 53%, with a balanced accuracy of 69%, which improves upon other methods significantly.
Conclusions:
The new tool will enable researchers to computationally identify nsSNVs that may affect drug responses. However, much larger training and testing datasets are needed to develop more reliable and accurate tools.
X-ray free electron lasers are used in measuring diffraction patterns from nanocrystals in the 'diffract-before-destroy' mode by outrunning radiation damage. The finite-sized nanocrystals provide an opportunity to recover intensity between Bragg spots by removing the modulating function that depends on crystal shape, i.e. the transform of the crystal shape. This shape-transform dividing-out scheme for solving the phase problem has been tested using simulated examples with cubic crystals. It provides a phasing method which does not require atomic resolution data, chemical modification to the sample, or modelling based on the protein databases. It is common to find multiple structural units (e.g. molecules, in symmetry-related positions) within a single unit cell, therefore incomplete unit cells (e.g. one additional molecule) can be observed at surface layers of crystals. In this work, the effects of such incomplete unit cells on the 'dividing-out' phasing algorithm are investigated using 2D crystals within the projection approximation. It is found that the incomplete unit cells do not hinder the recovery of the scattering pattern from a single unit cell (after dividing out the shape transforms from data merged from many nanocrystals of different sizes), assuming that certain unit-cell types are preferred. The results also suggest that the dynamic range of the data is a critical issue to be resolved in order to apply the shape transform method practically.
Serial femtosecond crystallography requires reliable and efficient delivery of fresh crystals across the beam of an X-ray free-electron laser over the course of an experiment. We introduce a double-flow focusing nozzle to meet this challenge, with significantly reduced sample consumption, while improving jet stability over previous generations of nozzles. We demonstrate its use to determine the first room-temperature structure of RNA polymerase II at high resolution, revealing new structural details. Moreover, the double flow-focusing nozzles were successfully tested with three other protein samples and the first room temperature structure of an extradiol ring-cleaving dioxygenase was solved by utilizing the improved operation and characteristics of these devices.
X-ray free-electron lasers provide novel opportunities to conduct single particle analysis on nanoscale particles. Coherent diffractive imaging experiments were performed at the Linac Coherent Light Source (LCLS), SLAC National Laboratory, exposing single inorganic core-shell nanoparticles to femtosecond hard-X-ray pulses. Each facetted nanoparticle consisted of a crystalline gold core and a differently shaped palladium shell. Scattered intensities were observed up to about 7 nm resolution. Analysis of the scattering patterns revealed the size distribution of the samples, which is consistent with that obtained from direct real-space imaging by electron microscopy. Scattering patterns resulting from single particles were selected and compiled into a dataset which can be valuable for algorithm developments in single particle scattering research.
