MOTIVATION: Untargeted metabolomic approaches hold a great promise as a diagnostic tool for inborn errors of metabolisms (IEMs) in the near future. However, the complexity of the involved data makes its application difficult and time consuming. Computational approaches, such as metabolic network simulations and machine learning, could significantly help to exploit metabolomic data to aid the diagnostic process. While the former suffers from limited predictive accuracy, the latter is normally able to generalize only to IEMs for which sufficient data are available. Here, we propose a hybrid approach that exploits the best of both worlds by building a mapping between simulated and real metabolic data through a novel method based on Siamese neural networks (SNN). RESULTS: The proposed SNN model is able to perform disease prioritization for the metabolic profiles of IEM patients even for diseases that it was not trained to identify. To the best of our knowledge, this has not been attempted before. The developed model is able to significantly outperform a baseline model that relies on metabolic simulations only. The prioritization performances demonstrate the feasibility of the method, suggesting that the integration of metabolic models and data could significantly aid the IEM diagnosis process in the near future. AVAILABILITY AND IMPLEMENTATION: Metabolic datasets used in this study are publicly available from the cited sources. The original data produced in this study, including the trained models and the simulated metabolic profiles, are also publicly available (Messa et al., 2020).

SUMMARY: Pathway-based expression profiles allow for high-level interpretation of transcriptomic data and systematic comparison of dysregulated cellular programs. We have previously demonstrated the efficacy of pathway-based approaches with two different applications: the Drug Set Enrichment Analysis and the Gene2drug analysis. Here we present a software tool that allows to easily convert gene-based profiles to pathway-based profiles and analyze them within the popular R framework. We also provide pre-computed profiles derived from the original Connectivity Map and its next generation release, i.e. the LINCS database. AVAILABILITY AND IMPLEMENTATION: the tool is implemented as the R/Bioconductor package gep2pep and can be freely downloaded from https://bioconductor.org/packages/gep2pep. SUPPLEMENTARY INFORMATION: Supplementary data are available at http://dsea.tigem.it/lincs.

Motivation: Drug repositioning has been proposed as an effective shortcut to drug discovery. The availability of large collections of transcriptional responses to drugs enables computational approaches to drug repositioning directly based on measured molecular effects. Results: We introduce a novel computational methodology for rational drug repositioning, which exploits the transcriptional responses following treatment with small molecule. Specifically, given a therapeutic target gene, a prioritization of potential effective drugs is obtained by assessing their impact on the transcription of genes in the pathway(s) including the target. We performed in silico validation and comparison with a state-of-art technique based on similar principles. We next performed experimental validation in two different real-case drug repositioning scenarios: (i) upregulation of the glutamate-pyruvate transaminase (GPT), which has been shown to induce reduction of oxalate levels in a mouse model of primary hyperoxaluria, and (ii) activation of the transcription factor TFEB, a master regulator of lysosomal biogenesis and autophagy, whose modulation may be beneficial in neurodegenerative disorders. Availability and implementation: A web tool for Gene2drug is freely available at http://gene2drug.tigem.it. An R package is under development and can be obtained from https://github.com/franapoli/gep2pep. Contact: [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

MOTIVATION: Automated screening approaches are able to rapidly identify a set of small molecules inducing a desired phenotype from large small-molecule libraries. However, the resulting set of candidate molecules is usually very diverse pharmacologically, thus little insight on the shared mechanism of action (MoA) underlying their efficacy can be gained. RESULTS: We introduce a computational method (Drug-Set Enrichment Analysis-DSEA) based on drug-induced gene expression profiles, which is able to identify the molecular pathways that are targeted by most of the drugs in the set. By diluting drug-specific effects unrelated to the phenotype of interest, DSEA is able to highlight phenotype-specific pathways, thus helping to formulate hypotheses on the MoA shared by the drugs in the set. We validated the method by analysing five different drug-sets related to well-known pharmacological classes. We then applied DSEA to identify the MoA shared by drugs known to be partially effective in rescuing mutant cystic fibrosis transmembrane conductance regulator (CFTR) gene function in Cystic Fibrosis. AVAILABILITY AND IMPLEMENTATION: The method is implemented as an online web tool publicly available at http://dsea.tigem.it. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Transcription factors (TFs) act downstream of the major signalling pathways functioning as master regulators of cell fate. Their activity is tightly regulated at the transcriptional, post-transcriptional and post-translational level. Proteins modifying TF activity are not easily identified by experimental high-throughput methods. RESULTS: We developed a computational strategy, called Differential Multi-Information (DMI), to infer post-translational modulators of a transcription factor from a compendium of gene expression profiles (GEPs). DMI is built on the hypothesis that the modulator of a TF (i.e. kinase/phosphatases), when expressed in the cell, will cause the TF target genes to be co-expressed. On the contrary, when the modulator is not expressed, the TF will be inactive resulting in a loss of co-regulation across its target genes. DMI detects the occurrence of changes in target gene co-regulation for each candidate modulator, using a measure called Multi-Information. We validated the DMI approach on a compendium of 5,372 GEPs showing its predictive ability in correctly identifying kinases regulating the activity of 14 different transcription factors. CONCLUSIONS: DMI can be used in combination with experimental approaches as high-throughput screening to efficiently improve both pathway and target discovery. An on-line web-tool enabling the user to use DMI to identify post-transcriptional modulators of a transcription factor of interest che be found at http://dmi.tigem.it.

SUMMARY: Elucidation of molecular targets of a compound [mode of action (MoA)] and its off-targets is a crucial step in drug development. We developed an online collaborative resource (MANTRA 2.0) that supports this process by exploiting similarities between drug-induced transcriptional profiles. Drugs are organized in a network of nodes (drugs) and edges (similarities) highlighting 'communities' of drugs sharing a similar MoA. A user can upload gene expression profiles before and after drug treatment in one or multiple cell types. An automated processing pipeline transforms the gene expression profiles into a unique drug 'node' embedded in the drug-network. Visual inspection of the neighbouring drugs and communities helps in revealing its MoA and to suggest new applications of known drugs (drug repurposing). MANTRA 2.0 allows storing and sharing user-generated network nodes, thus making MANTRA 2.0 a collaborative ever-growing resource. AVAILABILITY AND IMPLEMENTATION: The web tool is freely available for academic use at http://mantra.tigem.it.

We describe an innovative experimental and computational approach to control the expression of a protein in a population of yeast cells. We designed a simple control algorithm to automatically regulate the administration of inducer molecules to the cells by comparing the actual protein expression level in the cell population with the desired expression level. We then built an automated platform based on a microfluidic device, a time-lapse microscopy apparatus, and a set of motorized syringes, all controlled by a computer. We tested the platform to force yeast cells to express a desired fixed, or time-varying, amount of a reporter protein over thousands of minutes. The computer automatically switched the type of sugar administered to the cells, its concentration and its duration, according to the control algorithm. Our approach can be used to control expression of any protein, fused to a fluorescent reporter, provided that an external molecule known to (indirectly) affect its promoter activity is available.

MOTIVATION: Identification of differential expressed genes has led to countless new discoveries. However, differentially expressed genes are only a proxy for finding dysregulated pathways. The problem is to identify how the network of regulatory and physical interactions rewires in different conditions or in disease. RESULTS: We developed a procedure named DINA (DIfferential Network Analysis), which is able to identify set of genes, whose co-regulation is condition-specific, starting from a collection of condition-specific gene expression profiles. DINA is also able to predict which transcription factors (TFs) may be responsible for the pathway condition-specific co-regulation. We derived 30 tissue-specific gene networks in human and identified several metabolic pathways as the most differentially regulated across the tissues. We correctly identified TFs such as Nuclear Receptors as their main regulators and demonstrated that a gene with unknown function (YEATS2) acts as a negative regulator of hepatocyte metabolism. Finally, we showed that DINA can be used to make hypotheses on dysregulated pathways during disease progression. By analyzing gene expression profiles across primary and transformed hepatocytes, DINA identified hepatocarcinoma-specific metabolic and transcriptional pathway dysregulation. AVAILABILITY: We implemented an on-line web-tool http://dina.tigem.it enabling the user to apply DINA to identify tissue-specific pathways or gene signatures. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The connection between chromatin nuclear organization and gene activity is vividly illustrated by the observation that transcriptional coregulation of certain genes appears to be directly influenced by their spatial proximity. This fact poses the more general question of whether it is at all feasible that the numerous genes that are coregulated on a given chromosome, especially those at large genomic distances, might become proximate inside the nucleus. This problem is studied here using steered molecular dynamics simulations in order to enforce the colocalization of thousands of knowledge-based gene sequences on a model for the gene-rich human chromosome 19. Remarkably, it is found that most (≈ 88%) gene pairs can be brought simultaneously into contact. This is made possible by the low degree of intra-chromosome entanglement and the large number of cliques in the gene coregulatory network. A clique is a set of genes coregulated all together as a group. The constrained conformations for the model chromosome 19 are further shown to be organized in spatial macrodomains that are similar to those inferred from recent HiC measurements. The findings indicate that gene coregulation and colocalization are largely compatible and that this relationship can be exploited to draft the overall spatial organization of the chromosome in vivo. The more general validity and implications of these findings could be investigated by applying to other eukaryotic chromosomes the general and transferable computational strategy introduced here.

Regulation of gene expression is a carefully regulated phenomenon in the cell. “Reverse-engineering” algorithms try to reconstruct the regulatory interactions among genes from genome-scale measurements of gene expression profiles (microarrays). Mammalian cells express tens of thousands of genes; hence, hundreds of gene expression profiles are necessary in order to have acceptable statistical evidence of interactions between genes. As the number of profiles to be analyzed increases, so do computational costs and memory requirements. In this work, we designed and developed a parallel computing algorithm to reverse-engineer genome-scale gene regulatory networks from thousands of gene expression profiles. The algorithm is based on computing pairwise Mutual Information between each gene-pair. We successfully tested it to reverse engineer the Mus Musculus (mouse) gene regulatory network in liver from gene expression profiles collected from a public repository. A parallel hierarchical clustering algorithm was implemented to discover “communities” within the gene network. Network communities are enriched for genes involved in the same biological functions. The inferred network was used to identify two mitochondrial proteins.

Understanding the relationship between topology and dynamics of transcriptional regulatory networks in mammalian cells is essential to elucidate the biology of complex regulatory and signaling pathways. Here, we characterised, via a synthetic biology approach, a transcriptional positive feedback loop (PFL) by generating a clonal population of mammalian cells (CHO) carrying a stable integration of the construct. The PFL network consists of the Tetracycline-controlled transactivator (tTA), whose expression is regulated by a tTA responsive promoter (CMV-TET), thus giving rise to a positive feedback. The same CMV-TET promoter drives also the expression of a destabilised yellow fluorescent protein (d2EYFP), thus the dynamic behaviour can be followed by time-lapse microscopy. The PFL network was compared to an engineered version of the network lacking the positive feedback loop (NOPFL), by expressing the tTA mRNA from a constitutive promoter. Doxycycline was used to repress tTA activation (switch off), and the resulting changes in fluorescence intensity for both the PFL and NOPFL networks were followed for up to 43 h. We observed a striking difference in the dynamics of the PFL and NOPFL networks. Using non-linear dynamical models, able to recapitulate experimental observations, we demonstrated a link between network topology and network dynamics. Namely, transcriptional positive autoregulation can significantly slow down the "switch off" times, as compared to the non-autoregulated system. Doxycycline concentration can modulate the response times of the PFL, whereas the NOPFL always switches off with the same dynamics. Moreover, the PFL can exhibit bistability for a range of Doxycycline concentrations. Since the PFL motif is often found in naturally occurring transcriptional and signaling pathways, we believe our work can be instrumental to characterise their behaviour.

We investigated the possibility of gaining information on the mode of action of a set of compounds by means of Gene Ontology (GO) enrichment analysis. To this aim, we developed a new method, based on fuzzy-sets, which is able to compute sets of genes that are consistently differentially expressed when treating cells with the analyzed compounds. Then a Gene Ontology enrichment analysis is performed on these sets. The method has been tested on several different groups of drugs, whose similarity in mode of action has been predicted by a gene-expression based, unsupervised, approach and verified by searching literature. The obtained results show that GO terms that are overrepresented in these fuzzy sets provide a quick and "easy-tointerpret" view of the mode of action of the analyzed drugs.

MOTIVATION: The huge growth in gene expression data calls for the implementation of automatic tools for data processing and interpretation. RESULTS: We present a new and comprehensive machine learning data mining framework consisting in a non-linear PCA neural network for feature extraction, and probabilistic principal surfaces combined with an agglomerative approach based on Negentropy aimed at clustering gene microarray data. The method, which provides a user-friendly visualization interface, can work on noisy data with missing points and represents an automatic procedure to get, with no a priori assumptions, the number of clusters present in the data. Cell-cycle dataset and a detailed analysis confirm the biological nature of the most significant clusters. AVAILABILITY: The software described here is a subpackage part of the ASTRONEURAL package and is available upon request from the corresponding author. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Time series expression experiments are an increasingly popular method for studying a wide range of biological systems. Here we developed an algorithm that can infer the local network of gene-gene interactions surrounding a gene of interest. This is achieved by a perturbation of the gene of interest and subsequently measuring the gene expression profiles at multiple time points. We applied this algorithm to computer simulated data and to experimental data on a nine gene network in Escherichia coli. RESULTS: In this paper we show that it is possible to recover the gene regulatory network from a time series data of gene expression following a perturbation to the cell. We show this both on simulated data and on a nine gene subnetwork part of the DNA-damage response pathway (SOS pathway) in the bacteria E. coli. CONTACT: [email protected] SUPLEMENTARY INFORMATION: Supplementary data are available at http://dibernado.tigem.it

The recent technological advances are producing huge data sets in almost all fields of scientific research, from astronomy to genetics. Although each research field often requires ad-hoc, fine tuned, procedures to properly exploit all the available information inherently present in the data, there is an urgent need for a new generation of general computational theories and tools capable to boost most human activities of data analysis. Here, we propose probabilistic principal surfaces (PPS) as an effective high-D data visualization and clustering tool for data mining applications, emphasizing its flexibility and generality of use in data-rich field. In order to better illustrate the potentialities of the method, we also provide a real world case-study by discussing the use of PPS for the analysis of yeast gene expression levels from microarray chips.

Abstract Motivation: Structured non-coding RNAs (ncRNAs) have a very important functional role in the cell. No distinctive general features common to all ncRNA have yet been discovered. This makes it difficult to design computational tools able to detect novel ncRNAs in the genomic sequence. Results: We devised an algorithm able to detect conserved secondary structures in both pairwise and multiple DNA sequence alignments with computational time proportional to the square of the sequence length. We implemented the algorithm for the case of pairwise and three-way alignments and tested it on ncRNAs obtained from public databases. On the test sets, the pairwise algorithm has a specificity greater than 97% with a sensitivity varying from 22.26% for Blast alignments to 56.35% for structural alignments. The three-way algorithm behaves similarly. Our algorithm is able to efficiently detect a conserved secondary structure in multiple alignments. Availability: The executable file, the code and a PDF file containing the instructions, as well as the test data sets can be requested free of charge from the authors ([email protected]) or from the website http://www.tigem.it/research/dibernardo/default.htm