MOTIVATION: Generalizing machine learning models across small, high-dimensional, and heterogeneous biological datasets remains a critical challenge due to domain shifts caused by variations in data collection, population differences, and privacy constraints that restrict data sharing. Existing federated domain adaptation (FDA) approaches primarily rely on deep learning and focus on classification tasks, making them unsuitable for privacy-sensitive, small-scale regression problems in biomedical research. We introduce a privacy-preserving federated method for unsupervised domain adaptation in regression, enabling robust learning across distributed, high-dimensional datasets while maintaining full data privacy. RESULTS: Our method is the first to enable distributed training of Gaussian processes for domain adaptation, ensuring complete privacy through randomized encoding and secure aggregation. Unlike deep learning-based FDA approaches, our method is specifically designed for small-scale, high-dimensional biological data, overcoming prior limitations in scalability and generalization. We evaluate our approach on age prediction from DNA methylation data, demonstrating that it achieves performance comparable to non-private state-of-the-art methods while fully preserving data privacy. This work enables secure and effective cross-institutional collaboration in biomedical research without requiring raw data sharing. AVAILABILITY AND IMPLEMENTATION: The source code for our method is available at https://github.com/mdppml/FREDA.

Immunization through repeated direct venous inoculation of Plasmodium falciparum (Pf) sporozoites (PfSPZ) under chloroquine chemoprophylaxis, using the PfSPZ Chemoprophylaxis Vaccine (PfSPZ-CVac), induces high-level protection against controlled human malaria infection (CHMI). Humoral and cellular immunity contribute to vaccine efficacy but only limited information about the implicated Pf-specific antigens is available. Here, we examined Pf-specific antibody profiles, measured by protein arrays representing the full Pf proteome, of 40 placebo- and PfSPZ-immunized malaria-naïve volunteers from an earlier published PfSPZ-CVac dose-escalation trial. For this purpose, we both utilized and adapted supervised machine learning methods to identify predictive antibody profiles at two different time points: after immunization and before CHMI. We developed an adapted multitask support vector machine (SVM) approach and compared it to standard methods, i.e. single-task SVM, regularized logistic regression and random forests. Our results show, that the multitask SVM approach improved the classification performance to discriminate the protection status based on the underlying antibody-profiles while combining time- and dose-dependent data in the prediction model. Additionally, we developed the new fEature diStance exPlainabilitY (ESPY) method to quantify the impact of single antigens on the non-linear multitask SVM model and make it more interpretable. In conclusion, our multitask SVM model outperforms the studied standard approaches in regard of classification performance. Moreover, with our new explanation method ESPY, we were able to interpret the impact of Pf-specific antigen antibody responses that predict sterile protective immunity against CHMI after immunization. The identified Pf-specific antigens may contribute to a better understanding of immunity against human malaria and may foster vaccine development.

MOTIVATION: The size of available omics datasets is steadily increasing with technological advancement in recent years. While this increase in sample size can be used to improve the performance of relevant prediction tasks in healthcare, models that are optimized for large datasets usually operate as black boxes. In high-stakes scenarios, like healthcare, using a black-box model poses safety and security issues. Without an explanation about molecular factors and phenotypes that affected the prediction, healthcare providers are left with no choice but to blindly trust the models. We propose a new type of artificial neural network, named Convolutional Omics Kernel Network (COmic). By combining convolutional kernel networks with pathway-induced kernels, our method enables robust and interpretable end-to-end learning on omics datasets ranging in size from a few hundred to several hundreds of thousands of samples. Furthermore, COmic can be easily adapted to utilize multiomics data. RESULTS: We evaluated the performance capabilities of COmic on six different breast cancer cohorts. Additionally, we trained COmic models on multiomics data using the METABRIC cohort. Our models performed either better or similar to competitors on both tasks. We show how the use of pathway-induced Laplacian kernels opens the black-box nature of neural networks and results in intrinsically interpretable models that eliminate the need for post hoc explanation models. AVAILABILITY AND IMPLEMENTATION: Datasets, labels, and pathway-induced graph Laplacians used for the single-omics tasks can be downloaded at https://ibm.ent.box.com/s/ac2ilhyn7xjj27r0xiwtom4crccuobst/folder/48027287036. While datasets and graph Laplacians for the METABRIC cohort can be downloaded from the above mentioned repository, the labels have to be downloaded from cBioPortal at https://www.cbioportal.org/study/clinicalData?id=brca\_metabric. COmic source code as well as all scripts necessary to reproduce the experiments and analysis are publicly available at https://github.com/jditz/comics.

MOTIVATION: Machine learning methods can be used to support scientific discovery in healthcare-related research fields. However, these methods can only be reliably used if they can be trained on high-quality and curated datasets. Currently, no such dataset for the exploration of Plasmodium falciparum protein antigen candidates exists. The parasite P.falciparum causes the infectious disease malaria. Thus, identifying potential antigens is of utmost importance for the development of antimalarial drugs and vaccines. Since exploring antigen candidates experimentally is an expensive and time-consuming process, applying machine learning methods to support this process has the potential to accelerate the development of drugs and vaccines, which are needed for fighting and controlling malaria. RESULTS: We developed PlasmoFAB, a curated benchmark that can be used to train machine learning methods for the exploration of P.falciparum protein antigen candidates. We combined an extensive literature search with domain expertise to create high-quality labels for P.falciparum specific proteins that distinguish between antigen candidates and intracellular proteins. Additionally, we used our benchmark to compare different well-known prediction models and available protein localization prediction services on the task of identifying protein antigen candidates. We show that available general-purpose services are unable to provide sufficient performance on identifying protein antigen candidates and are outperformed by our models that were trained on this tailored data. AVAILABILITY AND IMPLEMENTATION: PlasmoFAB is publicly available on Zenodo with DOI 10.5281/zenodo.7433087. Furthermore, all scripts that were used in the creation of PlasmoFAB and the training and evaluation of machine learning models are open source and publicly available on GitHub here: https://github.com/msmdev/PlasmoFAB.

MOTIVATION: We present a multi-sequence generalization of Variational Information Bottleneck and call the resulting model Attentive Variational Information Bottleneck (AVIB). Our AVIB model leverages multi-head self-attention to implicitly approximate a posterior distribution over latent encodings conditioned on multiple input sequences. We apply AVIB to a fundamental immuno-oncology problem: predicting the interactions between T-cell receptors (TCRs) and peptides. RESULTS: Experimental results on various datasets show that AVIB significantly outperforms state-of-the-art methods for TCR-peptide interaction prediction. Additionally, we show that the latent posterior distribution learned by AVIB is particularly effective for the unsupervised detection of out-of-distribution amino acid sequences. AVAILABILITY AND IMPLEMENTATION: The code and the data used for this study are publicly available at: https://github.com/nec-research/vibtcr. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Missense variants in genes encoding ion channels are associated with a spectrum of severe diseases. Variant effects on biophysical function correlate with clinical features and can be categorized as gain- or loss-of-function. This information enables a timely diagnosis, facilitates precision therapy, and guides prognosis. Functional characterization presents a bottleneck in translational medicine. Machine learning models may be able to rapidly generate supporting evidence by predicting variant functional effects. Here, we describe a multi-task multi-kernel learning framework capable of harmonizing functional results and structural information with clinical phenotypes. This novel approach extends the human phenotype ontology towards kernel-based supervised machine learning. Our gain- or loss-of-function classifier achieves high performance (mean accuracy 0.853 SD 0.016, mean AU-ROC 0.912 SD 0.025), outperforming both conventional baseline and state-of-the-art methods. Performance is robust across different phenotypic similarity measures and largely insensitive to phenotypic noise or sparsity. Localized multi-kernel learning offered biological insight and interpretability by highlighting channels with implicit genotype-phenotype correlations or latent task similarity for downstream analysis.

The mechanisms triggering the human immunodeficiency virus type I (HIV-1) to switch the coreceptor usage from CCR5 to CXCR4 during the course of infection are not entirely understood. While low CD4+ T cell counts are associated with CXCR4 usage, a predominance of CXCR4 usage with still high CD4+ T cell counts remains puzzling. Here, we explore the hypothesis that viral adaptation to the human leukocyte antigen (HLA) complex, especially to the HLA class II alleles, contributes to the coreceptor switch. To this end, we sequence the viral gag and env protein with corresponding HLA class I and II alleles of a new cohort of 312 treatment-naive, subtype C, chronically-infected HIV-1 patients from South Africa. To estimate HLA adaptation, we develop a novel computational approach using Bayesian generalized linear mixed models (GLMMs). Our model allows to consider the entire HLA repertoire without restricting the model to pre-learned HLA-polymorphisms. In addition, we correct for phylogenetic relatedness of the viruses within the model itself to account for founder effects. Using our model, we observe that CXCR4-using variants are more adapted than CCR5-using variants (p-value = 1.34e-2). Additionally, adapted CCR5-using variants have a significantly lower predicted false positive rate (FPR) by the geno2pheno[coreceptor] tool compared to the non-adapted CCR5-using variants (p-value = 2.21e-2), where a low FPR is associated with CXCR4 usage. Consequently, estimating HLA adaptation can be an asset in predicting not only coreceptor usage, but also an approaching coreceptor switch in CCR5-using variants. We propose the usage of Bayesian GLMMs for modeling virus-host adaptation in general.

MOTIVATION: Diagnosis and treatment decisions on genomic data have become widespread as the cost of genome sequencing decreases gradually. In this context, disease-gene association studies are of great importance. However, genomic data are very sensitive when compared to other data types and contains information about individuals and their relatives. Many studies have shown that this information can be obtained from the query-response pairs on genomic databases. In this work, we propose a method that uses secure multi-party computation to query genomic databases in a privacy-protected manner. The proposed solution privately outsources genomic data from arbitrarily many sources to the two non-colluding proxies and allows genomic databases to be safely stored in semi-honest cloud environments. It provides data privacy, query privacy and output privacy by using XOR-based sharing and unlike previous solutions, it allows queries to run efficiently on hundreds of thousands of genomic data. RESULTS: We measure the performance of our solution with parameters similar to real-world applications. It is possible to query a genomic database with 3 000 000 variants with five genomic query predicates under 400 ms. Querying 1 048 576 genomes, each containing 1 000 000 variants, for the presence of five different query variants can be achieved approximately in 6 min with a small amount of dedicated hardware and connectivity. These execution times are in the right range to enable real-world applications in medical research and healthcare. Unlike previous studies, it is possible to query multiple databases with response times fast enough for practical application. To the best of our knowledge, this is the first solution that provides this performance for querying large-scale genomic data. AVAILABILITY AND IMPLEMENTATION: https://gitlab.com/DIFUTURE/privacy-preserving-variant-queries. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Training sophisticated machine learning models usually requires many training samples. Especially in healthcare settings these samples can be very expensive, meaning that one institution alone usually does not have enough. Merging privacy-sensitive data from different sources is usually restricted by data security and data protection measures. This can lead to approaches that reduce data quality by putting noise onto the variables (e.g., in epsilon-differential privacy) or omitting certain values (e.g., for k-anonymity). Other measures based on cryptographic methods can lead to very time-consuming computations, which is especially problematic for larger multi-omics data. We address this problem by introducing ESCAPED, which stands for Efficient SeCure And PrivatE Dot product framework. ESCAPED enables the computation of the dot product of vectors from multiple sources on a third-party, which later trains kernel-based machine learning algorithms, while neither sacrificing privacy nor adding noise. We have evaluated our framework on drug resistance prediction for HIV-infected people and multi-omics dimensionality reduction and clustering problems in precision medicine. In terms of execution time, our framework significantly outperforms the best-fitting existing approaches without sacrificing the performance of the algorithm. Even though we only present the benefit for kernel-based algorithms, our framework can open up new research opportunities for further machine learning models that require the dot product of vectors from multiple sources.

MOTIVATION: The use of genome data for diagnosis and treatment is becoming increasingly common. Researchers need access to as many genomes as possible to interpret the patient genome, to obtain some statistical patterns and to reveal disease-gene relationships. The sensitive information contained in the genome data and the high risk of re-identification increase the privacy and security concerns associated with sharing such data. In this article, we present an approach to identify disease-associated variants and genes while ensuring patient privacy. The proposed method uses secure multi-party computation to find disease-causing mutations under specific inheritance models without sacrificing the privacy of individuals. It discloses only variants or genes obtained as a result of the analysis. Thus, the vast majority of patient data can be kept private. RESULTS: Our prototype implementation performs analyses on thousands of genomic data in milliseconds, and the runtime scales logarithmically with the number of patients. We present the first inheritance model (recessive, dominant and compound heterozygous) based privacy-preserving analyses of genomic data to find disease-causing mutations. Furthermore, we re-implement the privacy-preserving methods (MAX, SETDIFF and INTERSECTION) proposed in a previous study. Our MAX, SETDIFF and INTERSECTION implementations are 2.5, 1122 and 341 times faster than the corresponding operations of the state-of-the-art protocol, respectively. AVAILABILITY AND IMPLEMENTATION: https://gitlab.com/DIFUTURE/privacy-preserving-genomic-diagnosis. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Predictive models are a powerful tool for solving complex problems in computational biology. They are typically designed to predict or classify data coming from the same unknown distribution as the training data. In many real-world settings, however, uncontrolled biological or technical factors can lead to a distribution mismatch between datasets acquired at different times, causing model performance to deteriorate on new data. A common additional obstacle in computational biology is scarce data with many more features than samples. To address these problems, we propose a method for unsupervised domain adaptation that is based on a weighted elastic net. The key idea of our approach is to compare dependencies between inputs in training and test data and to increase the cost of differently behaving features in the elastic net regularization term. In doing so, we encourage the model to assign a higher importance to features that are robust and behave similarly across domains. RESULTS: We evaluate our method both on simulated data with varying degrees of distribution mismatch and on real data, considering the problem of age prediction based on DNA methylation data across multiple tissues. Compared with a non-adaptive standard model, our approach substantially reduces errors on samples with a mismatched distribution. On real data, we achieve far lower errors on cerebellum samples, a tissue which is not part of the training data and poorly predicted by standard models. Our results demonstrate that unsupervised domain adaptation is possible for applications in computational biology, even with many more features than samples. AVAILABILITY AND IMPLEMENTATION: Source code is available at https://github.com/PfeiferLabTue/wenda. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Most string kernels for comparison of genomic sequences are generally tied to using (absolute) positional information of the features in the individual sequences. This poses limitations when comparing variable-length sequences using such string kernels. For example, profiling chromatin interactions by 3C-based experiments results in variable-length genomic sequences (restriction fragments). Here, exact position-wise occurrence of signals in sequences may not be as important as in the scenario of analysis of the promoter sequences, that typically have a transcription start site as reference. Existing position-aware string kernels have been shown to be useful for the latter scenario. In this work, we propose a novel approach for sequence comparison that enables larger positional freedom than most of the existing approaches, can identify a possibly dispersed set of features in comparing variable-length sequences, and can handle both the aforementioned scenarios. Our approach, \emph{CoMIK}, identifies not just the features useful towards classification but also their locations in the variable-length sequences, as evidenced by the results of three binary classification experiments, aided by recently introduced visualization techniques. Furthermore, we show that we are able to efficiently retrieve and interpret the weight vector for the complex setting of multiple multi-instance kernels.

BACKGROUND: Knowing the three-dimensional (3D) structure of the chromatin is important for obtaining a complete picture of the regulatory landscape. Changes in the 3D structure have been implicated in diseases. While there exist approaches that attempt to predict the long-range chromatin interactions, they focus only on interactions between specific genomic regions - the promoters and enhancers, neglecting other possibilities, for instance, the so-called structural interactions involving intervening chromatin. RESULTS: We present a method that can be trained on 5C data using the genetic sequence of the candidate loci to predict potential genome-wide interaction partners of a particular locus of interest. We have built locus-specific support vector machine (SVM)-based predictors using the oligomer distance histograms (ODH) representation. The method shows good performance with a mean test AUC (area under the receiver operating characteristic (ROC) curve) of 0.7 or higher for various regions across cell lines GM12878, K562 and HeLa-S3. In cases where any locus did not have sufficient candidate interaction partners for model training, we employed multitask learning to share knowledge between models of different loci. In this scenario, across the three cell lines, the method attained an average performance increase of 0.09 in the AUC. Performance evaluation of the models trained on 5C data regarding prediction on an independent high-resolution Hi-C dataset (which is a rather hard problem) shows 0.56 AUC, on average. Additionally, we have developed new, intuitive visualization methods that enable interpretation of sequence signals that contributed towards prediction of locus-specific interaction partners. The analysis of these sequence signals suggests a potential general role of short tandem repeat sequences in genome organization. CONCLUSIONS: We demonstrated how our approach can 1) provide insights into sequence features of locus-specific interaction partners, and 2) also identify their cell-line specificity. That our models deem short tandem repeat sequences as discriminative for prediction of potential interaction partners, suggests that they could play a larger role in genome organization. Thus, our approach can (a) be beneficial to broadly understand, at the sequence-level, chromatin interactions and higher-order structures like (meta-) topologically associating domains (TADs); (b) study regions omitted from existing prediction approaches using various information sources (e.g., epigenetic information); and

Treatment with broadly neutralizing antibodies (bNAbs) has proven effective against HIV-1 infections in humanized mice, non-human primates, and humans. Due to the high mutation rate of HIV-1, resistance testing of the patient's viral strains to the bNAbs is still inevitable. So far, bNAb resistance can only be tested in expensive and time-consuming neutralization experiments. Here, we introduce well-performing computational models that predict the neutralization response of HIV-1 to bNAbs given only the envelope sequence of the virus. Using non-linear support vector machines based on a string kernel, the models learnt even the important binding sites of bNAbs with more complex epitopes, i.e., the CD4 binding site targeting bNAbs, proving thereby the biological relevance of the models. To increase the interpretability of the models, we additionally provide a new kind of motif logo for each query sequence, visualizing those residues of the test sequence that influenced the prediction outcome the most. Moreover, we predicted the neutralization sensitivity of around 34,000 HIV-1 samples from different time points to a broad range of bNAbs, enabling the first analysis of HIV resistance to bNAbs on a global scale. The analysis showed for many of the bNAbs a trend towards antibody resistance over time, which had previously only been discovered for a small non-representative subset of the global HIV-1 population.

MOTIVATION: Despite ongoing cancer research, available therapies are still limited in quantity and effectiveness, and making treatment decisions for individual patients remains a hard problem. Established subtypes, which help guide these decisions, are mainly based on individual data types. However, the analysis of multidimensional patient data involving the measurements of various molecular features could reveal intrinsic characteristics of the tumor. Large-scale projects accumulate this kind of data for various cancer types, but we still lack the computational methods to reliably integrate this information in a meaningful manner. Therefore, we apply and extend current multiple kernel learning for dimensionality reduction approaches. On the one hand, we add a regularization term to avoid overfitting during the optimization procedure, and on the other hand, we show that one can even use several kernels per data type and thereby alleviate the user from having to choose the best kernel functions and kernel parameters for each data type beforehand. RESULTS: We have identified biologically meaningful subgroups for five different cancer types. Survival analysis has revealed significant differences between the survival times of the identified subtypes, with P values comparable or even better than state-of-the-art methods. Moreover, our resulting subtypes reflect combined patterns from the different data sources, and we demonstrate that input kernel matrices with only little information have less impact on the integrated kernel matrix. Our subtypes show different responses to specific therapies, which could eventually assist in treatment decision making. AVAILABILITY AND IMPLEMENTATION: An executable is available upon request.

DNA methylation is an epigenetic modification known to play a prime role in gene silencing and is an important topic in epigenetic research. However, due to technology-dependent errors there are inconsistencies between methylation measurements from different methods [ 1 ]. Incorrect methylation calls could result in the discovery of spurious associations between methylation patterns and specific phenotypes in epigenome-wide association studies (EWAS). We worked towards assigning a measure of confidence to individual CpGs to down-weigh or exclude positions with inconsistent measurements in such studies. We used methylation measurements from the Infinium HumanMethylation450 microarray (β450K) and whole genome bisulfite sequencing (βWGBS) to evaluate whether locus-specific measurement differences, Δβ = β450K − βWGBS, are predictable using statistical learning techniques. Methylation for Illumina WGBS data from HepaRGd7R2 was called with Bis-SNP [ 2 ], while methylation for Infinium 450K data from the same cell line was determined using RnBeads [ 3 ] and normalized with BMIQ [ 4 ]. For a uniform feature representation, we considered windows of reads overlapping with CpGs on the microarray (Figure 1 ). As predictors we examined sets of read sequences, their consensus sequences (with and without base frequencies), and non-sequence features such as base quality and depth of coverage. To obtain a predictive model independent of the methylation state, we masked CpG positions by introducing gaps or zeroing base frequencies. Data preprocessing . (1) Only reads overlapping with a CpG on the Infinium 450K chip are retained. (2) Windows are extended to the left and right of each CpG according to the maximum read length, yielding a uniform feature representation. (3) For each CpG, a consensus sequence is formed from its corresponding set of reads. Additionally, the position-specific frequency of each base is extracted. (4) Finally, CpG positions are masked by introducing gaps in the sequence or zeroing frequencies. To predict Δβ, we built support vector regression models based on Illumina WGBS data. Read similarity was measured with numerical, string [ 5 – 7 ], and set kernels [ 8 ]. We introduced the notion of hybrid string kernels to afford a similarity measure for both numeric and string input simultaneously. These kernels are based on scaling the motif similarity scores of two sequences according to the similarity of their base frequency profiles. For a read-based set kernel utilizing the weighted degree kernel with shifts [ 6 ], we found that the predicted values of Δβ correlated significantly with the observed outcomes (r = 0.37, p-value < 2.2 · 10−16). Furthermore, the hybrid weighted degree kernel (r = 0.234) outperformed the weighted degree kernel with shifts (r = 0.22) by also considering the frequencies of individual bases in addition to the consensus sequences. Non-sequence features were less predictive of the outcome than the sequence, e.g., RBF kernels on base quality and depth of coverage attained only correlations of r = 0.057 and r = 0.003 with the outcome, respectively. To our knowledge, this is the first approach indicating that differences between methylation measurements from bisulfite sequencing and the Infinium HumanMethylation450 microarray are predictable from the reads. The results suggest that features beside the sequence play only a minuscule role in the emergence of inconsistent methylation measurements. We were able to show that, in this scenario, set kernels and hybrid string kernels provide well-suited similarity measures. Further work is necessary to validate the model's generalizability for data from other cell lines and to evaluate its practical merit.

MOTIVATION: Due to the high mutation rate of human immunodeficiency virus (HIV), drug-resistant-variants emerge frequently. Therefore, researchers are constantly searching for new ways to attack the virus. One new class of anti-HIV drugs is the class of coreceptor antagonists that block cell entry by occupying a coreceptor on CD4 cells. This type of drug just has an effect on the subset of HIVs that use the inhibited coreceptor. A good prediction of whether the viral population inside a patient is susceptible to the treatment is hence very important for therapy decisions and pre-requisite to administering the respective drug. The first prediction models were based on data from Sanger sequencing of the V3 loop of HIV. Recently, a method based on next-generation sequencing (NGS) data was introduced that predicts labels for each read separately and decides on the patient label through a percentage threshold for the resistant viral minority. RESULTS: We model the prediction problem on the patient level taking the information of all reads from NGS data jointly into account. This enables us to improve prediction performance for NGS data, but we can also use the trained model to improve predictions based on Sanger sequencing data. Therefore, also laboratories without NGS capabilities can benefit from the improvements. Furthermore, we show which amino acids at which position are important for prediction success, giving clues on how the interaction mechanism between the V3 loop and the particular coreceptors might be influenced. AVAILABILITY: A webserver is available at http://coreceptor.bioinf.mpi-inf.mpg.de. CONTACT: [email protected].

BACKGROUND: Mass Spectrometry coupled to Liquid Chromatography (LC-MS) is commonly used to analyze the protein content of biological samples in large scale studies. The data resulting from an LC-MS experiment is huge, highly complex and noisy. Accordingly, it has sparked new developments in Bioinformatics, especially in the fields of algorithm development, statistics and software engineering. In a quantitative label-free mass spectrometry experiment, crucial steps are the detection of peptide features in the mass spectra and the alignment of samples by correcting for shifts in retention time. At the moment, it is difficult to compare the plethora of algorithms for these tasks. So far, curated benchmark data exists only for peptide identification algorithms but no data that represents a ground truth for the evaluation of feature detection, alignment and filtering algorithms. RESULTS: We present LC-MSsim, a simulation software for LC-ESI-MS experiments. It simulates ESI spectra on the MS level. It reads a list of proteins from a FASTA file and digests the protein mixture using a user-defined enzyme. The software creates an LC-MS data set using a predictor for the retention time of the peptides and a model for peak shapes and elution profiles of the mass spectral peaks. Our software also offers the possibility to add contaminants, to change the background noise level and includes a model for the detectability of peptides in mass spectra. After the simulation, LC-MSsim writes the simulated data to mzData, a public XML format. The software also stores the positions (monoisotopic m/z and retention time) and ion counts of the simulated ions in separate files. CONCLUSION: LC-MSsim generates simulated LC-MS data sets and incorporates models for peak shapes and contaminations. Algorithm developers can match the results of feature detection and alignment algorithms against the simulated ion lists and meaningful error rates can be computed. We anticipate that LC-MSsim will be useful to the wider community to perform benchmark studies and comparisons between computational tools.

BACKGROUND: Mass spectrometry is an essential analytical technique for high-throughput analysis in proteomics and metabolomics. The development of new separation techniques, precise mass analyzers and experimental protocols is a very active field of research. This leads to more complex experimental setups yielding ever increasing amounts of data. Consequently, analysis of the data is currently often the bottleneck for experimental studies. Although software tools for many data analysis tasks are available today, they are often hard to combine with each other or not flexible enough to allow for rapid prototyping of a new analysis workflow. RESULTS: We present OpenMS, a software framework for rapid application development in mass spectrometry. OpenMS has been designed to be portable, easy-to-use and robust while offering a rich functionality ranging from basic data structures to sophisticated algorithms for data analysis. This has already been demonstrated in several studies. CONCLUSION: OpenMS is available under the Lesser GNU Public License (LGPL) from the project website at http://www.openms.de.

MOTIVATION: Experimental techniques in proteomics have seen rapid development over the last few years. Volume and complexity of the data have both been growing at a similar rate. Accordingly, data management and analysis are one of the major challenges in proteomics. Flexible algorithms are required to handle changing experimental setups and to assist in developing and validating new methods. In order to facilitate these studies, it would be desirable to have a flexible 'toolbox' of versatile and user-friendly applications allowing for rapid construction of computational workflows in proteomics. RESULTS: We describe a set of tools for proteomics data analysis-TOPP, The OpenMS Proteomics Pipeline. TOPP provides a set of computational tools which can be easily combined into analysis pipelines even by non-experts and can be used in proteomics workflows. These applications range from useful utilities (file format conversion, peak picking) over wrapper applications for known applications (e.g. Mascot) to completely new algorithmic techniques for data reduction and data analysis. We anticipate that TOPP will greatly facilitate rapid prototyping of proteomics data evaluation pipelines. As such, we describe the basic concepts and the current abilities of TOPP and illustrate these concepts in the context of two example applications: the identification of peptides from a raw dataset through database search and the complex analysis of a standard addition experiment for the absolute quantitation of biomarkers. The latter example demonstrates TOPP's ability to construct flexible analysis pipelines in support of complex experimental setups. AVAILABILITY: The TOPP components are available as open-source software under the lesser GNU public license (LGPL). Source code is available from the project website at www.OpenMS.de

BACKGROUND: High-throughput peptide and protein identification technologies have benefited tremendously from strategies based on tandem mass spectrometry (MS/MS) in combination with database searching algorithms. A major problem with existing methods lies within the significant number of false positive and false negative annotations. So far, standard algorithms for protein identification do not use the information gained from separation processes usually involved in peptide analysis, such as retention time information, which are readily available from chromatographic separation of the sample. Identification can thus be improved by comparing measured retention times to predicted retention times. Current prediction models are derived from a set of measured test analytes but they usually require large amounts of training data. RESULTS: We introduce a new kernel function which can be applied in combination with support vector machines to a wide range of computational proteomics problems. We show the performance of this new approach by applying it to the prediction of peptide adsorption/elution behavior in strong anion-exchange solid-phase extraction (SAX-SPE) and ion-pair reversed-phase high-performance liquid chromatography (IP-RP-HPLC). Furthermore, the predicted retention times are used to improve spectrum identifications by a p-value-based filtering approach. The approach was tested on a number of different datasets and shows excellent performance while requiring only very small training sets (about 40 peptides instead of thousands). Using the retention time predictor in our retention time filter improves the fraction of correctly identified peptide mass spectra significantly. CONCLUSION: The proposed kernel function is well-suited for the prediction of chromatographic separation in computational proteomics and requires only a limited amount of training data. The performance of this new method is demonstrated by applying it to peptide retention time prediction in IP-RP-HPLC and prediction of peptide sample fractionation in SAX-SPE. Finally, we incorporate the predicted chromatographic behavior in a p-value based filter to improve peptide identifications based on liquid chromatography-tandem mass spectrometry.

Biological data mining using kernel methods can be improved by a task-specific choice of the kernel function. Oligo kernels for genomic sequence analysis have proven to have a high discriminative power and to provide interpretable results. Oligo kernels that consider subsequences of different lengths can be combined and parameterized to increase their flexibility. For adapting these parameters efficiently, gradient-based optimization of the kernel-target alignment is proposed. The power of this new, general model selection procedure and the benefits of fitting kernels to problem classes are demonstrated by adapting oligo kernels for bacterial gene start detection.

UNLABELLED: We provide the tool 'TICO' (Translation Initiation site COrrection) for improving the results of conventional gene finders for prokaryotic genomes with regard to exact localization of the translation initiation site (TIS). At the current state TICO provides an interface for direct post processing of the predictions obtained from the widely used program GLIMMER. Our program is based on a clustering algorithm for completely unsupervised scoring of potential TIS locations. AVAILABILITY: Our tool can be freely accessed through a web interface at http://tico.gobics.de/ CONTACT: [email protected]