Recent advancements in self-supervised learning in the point cloud domain have demonstrated significant potential. However, these methods often suffer from drawbacks such as lengthy pre-training time, the necessity of reconstruction in the input space, and the necessity of additional modalities. In order to address these issues, we introduce Point-JEPA, a joint embedding predictive architecture designed specifically for point cloud data. To this end, we introduce a sequencer that orders point cloud patch embeddings to efficiently compute and utilize their proximity based on their indices during target and context selection. The sequencer also allows shared computations of the patch embeddings' proximity between context and target selection, further improving the efficiency. Experimentally, our method demonstrates state-of-the-art performance while avoiding the re-construction in the input space or additional modality. In particular, Point-JEPA attains a classification accuracy of 93.7 ±0.2 % for linear SVM on ModelNet40 surpassing all other self-supervised models. Moreover, Point-JEPA also establishes new state-of-the-art performance levels across all four few-shot learning evaluation frameworks. The code is available at https://github.com/Ayumu-J-S/Point-JEPA

Probabilistic programming languages are used to write probabilistic models to make probabilistic inferences. A number of rigorous semantics have recently been proposed that are now available to carry out formal verification of probabilistic programs. In this paper, we extend an existing formalization of measure and integration theory with s-finite kernels, a mathematical structure to interpret typing judgments in the semantics of a probabilistic programming language. The resulting library makes it possible to reason formally about transformations of probabilistic programs and their execution.

UNLABELLED: XiP (eXtensible integrative Pipeline) is a flexible, editable and modular environment with a user-friendly interface that does not require previous advanced programming skills to run, construct and edit workflows. XiP allows the construction of workflows by linking components written in both R and Java, the analysis of high-throughput data in grid engine systems and also the development of customized pipelines that can be encapsulated in a package and distributed. XiP already comes with several ready-to-use pipeline flows for the most common genomic and transcriptomic analysis and ∼300 computational components. AVAILABILITY: XiP is open source, freely available under the Lesser General Public License (LGPL) and can be downloaded from http://xip.hgc.jp.

UNLABELLED: Protein-protein interactions (PPIs) are mediated through specific regions on proteins. Some proteins have two or more protein interacting regions (IRs) and some IRs are competitively used for interactions with different proteins. IRView currently contains data for 3417 IRs in human and mouse proteins. The data were obtained from different sources and combined with annotated region data from InterPro. Information on non-synonymous single nucleotide polymorphism sites and variable regions owing to alternative mRNA splicing is also included. The IRView web interface displays all IR data, including user-uploaded data, on reference sequences so that the positional relationship between IRs can be easily understood. IRView should be useful for analyzing underlying relationships between the proteins behind the PPI networks. AVAILABILITY: IRView is publicly available on the web at http://ir.hgc.jp/

SUMMARY: Manual curation and validation of large-scale biological pathways are required to obtain high-quality pathway databases. In a typical curation process, model validation and model update based on appropriate feedback are repeated and requires considerable cooperation of scientists. We have developed a CSO (Cell System Ontology) validator to reduce the repetition and time during the curation process. This tool assists in quickly obtaining agreement among curators and domain experts and in providing a consistent and accurate pathway database. AVAILABILITY: The tool is available on http://csovalidator.csml.org. CONTACT: [email protected].

UNLABELLED: Model checking is playing an increasingly important role in systems biology as larger and more complex biological pathways are being modeled. In this article we report the release of an efficient model checker MIRACH 1.0, which supports any model written in popular formats such as CSML and SBML. MIRACH is integrated with a Petri-net-based simulation engine, enabling efficient online (on-the-fly) checking. In our experiment, by using Levchenko et al. model, we reveal that timesaving gains by using MIRACH easily surpass 400% compared with its offline-based counterpart. AVAILABILITY AND IMPLEMENTATION: MIRACH 1.0 was developed using Java and thus executable on any platform installed with JDK 6.0 (not JRE 6.0) or later. MIRACH 1.0, along with its source codes, documentation and examples are available at http://sourceforge.net/projects/mirach/ under the LGPLv3 license.

SUMMARY: The Macrophage Pathway Knowledgebase (MACPAK) is a computational system that allows biomedical researchers to query and study the dynamic behaviors of macrophage molecular pathways. It integrates the knowledge of 230 reviews that were carefully checked by specialists for their accuracy and then converted to 230 dynamic mathematical pathway models. MACPAK comprises a total of 24 009 entities and 12 774 processes and is described in the Cell System Markup Language (CSML), an XML format that runs on the Cell Illustrator platform and can be visualized with a customized Cytoscape for further analysis. AVAILABILITY: MACPAK can be accessed via an interactive web site at http://macpak.csml.org. The CSML pathway models are available under the Creative Commons license.

SUMMARY: Data assimilation (DA) is a computational approach that estimates unknown parameters in a pathway model using time-course information. Particle filtering, the underlying method used, is a well-established statistical method that approximates the joint posterior distributions of parameters by using sequentially generated Monte Carlo samples. In this article, we report the release of Java-based software (DA 1.0) with an intuitive and user-friendly interface to allow users to carry out parameters estimation using DA. AVAILABILITY AND IMPLEMENTATION: DA 1.0 was developed using Java and thus would be executable on any platform installed with JDK 6.0 (not JRE 6.0) or later. DA 1.0 is freely available for academic users and can be launched or downloaded from http://da.csml.org.

MOTIVATION: The recent metagenome analysis has been producing a large number of host-unassigned viruses. Although assigning viruses to their hosts is basically important not only for virology but also for prevention of epidemic, it has been a laborious and difficult task to date. The only effective method for this purpose has been to find them in a same microscopic view. Now, we tried a computational approach based on genome sequences of bacteria and phages, introducing a physicochemical parameter, SOSS (set of oligostickiness similarity score) derived from oligostickiness, a measure of binding affinity of oligonucleotides to template DNA. RESULTS: We could confirm host-parasite relationships of bacteria and their phages by SOSS analysis: all phages tested (25 species) had a remarkably higher SOSS value with its host than with unrelated bacteria. Interestingly, according to SOSS values, lysogenic phages such as lambda phage (host: Escherichia coli) or SPP1 (host: Bacillus subtilis) have distinctively higher similarity with its host than its non-lysogenic (excretive or virulent) ones such as fd and T4 (host: E.coli) or phages gamma and PZA (host: B.subtilis). This finding is very promising for assigning host-unknown viruses to its host. We also investigated the relationship in codon usage frequency or G+C content of genomes to interpret the phenomenon revealed by SOSS analysis, obtaining evidences which support the hypothesis that higher SOSS values resulted from the cohabitation in the same environment which may cause the common biased mutation. Thus, lysogenic phages which stay inside longer resemble the host.

BACKGROUND: Some splicing isoform-specific transcriptional regulations are related to disease. Therefore, detection of disease specific splice variations is the first step for finding disease specific transcriptional regulations. Affymetrix Human Exon 1.0 ST Array can measure exon-level expression profiles that are suitable to find differentially expressed exons in genome-wide scale. However, exon array produces massive datasets that are more than we can handle and analyze on personal computer. RESULTS: We have developed ExonMiner that is the first all-in-one web service for analysis of exon array data to detect transcripts that have significantly different splicing patterns in two cells, e.g. normal and cancer cells. ExonMiner can perform the following analyses: (1) data normalization, (2) statistical analysis based on two-way ANOVA, (3) finding transcripts with significantly different splice patterns, (4) efficient visualization based on heatmaps and barplots, and (5) meta-analysis to detect exon level biomarkers. We implemented ExonMiner on a supercomputer system in order to perform genome-wide analysis for more than 300,000 transcripts in exon array data, which has the potential to reveal the aberrant splice variations in cancer cells as exon level biomarkers. CONCLUSION: ExonMiner is well suited for analysis of exon array data and does not require any installation of software except for internet browsers. What all users need to do is to access the ExonMiner URL http://ae.hgc.jp/exonminer. Users can analyze full dataset of exon array data within hours by high-level statistical analysis with sound theoretical basis that finds aberrant splice variants as biomarkers.

BACKGROUND: Comprehensive description of the behavior of cellular components in a quantitative manner is essential for systematic understanding of biological events. Recent LC-MS/MS (tandem mass spectrometry coupled with liquid chromatography) technology, in combination with the SILAC (Stable Isotope Labeling by Amino acids in Cell culture) method, has enabled us to make relative quantitation at the proteome level. The recent report by Blagoev et al. (Nat. Biotechnol., 22, 1139-1145, 2004) indicated that this method was also applicable for the time-course analysis of cellular signaling events. Relative quatitation can easily be performed by calculating the ratio of peak intensities corresponding to differentially labeled peptides in the MS spectrum. As currently available software requires some GUI applications and is time-consuming, it is not suitable for processing large-scale proteome data. RESULTS: To resolve this difficulty, we developed an algorithm that automatically detects the peaks in each spectrum. Using this algorithm, we developed a software tool named AYUMS that automatically identifies the peaks corresponding to differentially labeled peptides, compares these peaks, calculates each of the peak ratios in mixed samples, and integrates them into one data sheet. This software has enabled us to dramatically save time for generation of the final report. CONCLUSION: AYUMS is a useful software tool for comprehensive quantitation of the proteome data generated by LC-MS/MS analysis. This software was developed using Java and runs on Linux, Windows, and Mac OS X. Please contact [email protected] if you are interested in the application. The project web page is http://www.csml.org/ayums/.

MOTIVATION: An unmanageably large amount of data on genome sequences is accumulating, prompting researchers to develop new methods to analyze them. We have devised a novel method designated oligostickiness, a measure roughly proportional to the binding affinity of an oligonucleotide to a DNA of interest, in order to analyze genome sequences as a whole. RESULTS: Fifteen representative genomes such as Bacillus subtilis, Escherichia coli, Saccharomyces cerevisiae, Caenorhabditis elegans, H. sapiens and others were analyzed by this method using more than 50 probe dodecanucleotides, offering the following findings: (i) Genome sequences can be specifically featured by way of oligostickiness maps. (ii) Oligostickiness analysis, which is similar to but more informative than (G + C) content or repetitive sequence analysis, can reveal intra-genomic structures such as mosaic structures (E. coli and B. subtilis) and highly sticky/non-sticky regions of biological meanings. (iii) Some probe oligonucleotides such as dC(12) and dT(12) can be used for classifying genomes, clearly discriminating prokaryotes and eukaryotes. (iv) Based on global oligostickiness, which is the average value of the local oligostickinesses, the features of a genome could be visualized in spider web mode. The pattern of a spider web as well as a set of oligostickiness maps is highly characteristic to each genome or chromosome. Thus, we called it as chromosome texture, leading to a finding that all the chromosomes contained in a cell, so far investigated, have a common texture. AVAILABILITY: Oligostickinesses maps used in this work are available at http://gp.fms.saitama-u.ac.jp/ CONTACT: [email protected]