Motivation: Secondary metabolites (SM) are structurally diverse natural products of high pharmaceutical importance. Genes involved in their biosynthesis are often organized in clusters, i.e., are co-localized and co-expressed. In silico cluster prediction in eukaryotic genomes remains problematic mainly due to the high variability of the clusters’ content and lack of other distinguishing sequence features. Results: We present Cluster Assignment by Islands of Sites (CASSIS), a method for SM cluster prediction in eukaryotic genomes, and Secondary Metabolites by InterProScan (SMIPS), a tool for genome-wide detection of SM key enzymes (‘anchor’ genes): polyketide synthases, non-ribosomal peptide synthetases and dimethylallyl tryptophan synthases. Unlike other tools based on protein similarity, CASSIS exploits the idea of co-regulation of the cluster genes, which assumes the existence of common regulatory patterns in the cluster promoters. The method searches for ‘islands’ of enriched cluster-specific motifs in the vicinity of anchor genes. It was validated in a series of cross-validation experiments and showed high sensitivity and specificity. Availability and implementation: CASSIS and SMIPS are freely available at https://sbi.hki-jena.de/cassis. Contact: thomas.wolf@leibniz-hki.de or ekaterina.shelest@leibniz-hki.de Supplementary information: Supplementary data are available at Bioinformatics online.

The immune system is considered a key driver of atherosclerosis, and beyond proteins and microRNAs (miRs), long non-coding RNAs (lncRNAs) are implicated in immune control. We previously described that lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is involved in cardiac innate immunity in a myocarditis model. Here, we investigated the impact of MALAT1 deficiency upon atherosclerosis development. Heterozygous MALAT1-deficient ApoE-/- mice displayed massive immune system dysregulation and atherosclerosis within 2 months even when kept on normal diet. Aortic plaque area (P < 0.05) and aortic root plaque size (P < 0.001) were increased in MALAT1-deficient vs. MALAT1-wildtype ApoE-/- mice. Serum levels of interferon-γ (IFN-γ), tumour necrosis factor (TNF), and interleukin 6 (IL6) were elevated (P < 0.001) in MALAT1-deficient animals. MALAT1-deficient bone marrow-derived macrophages showed enhanced expression of TNF (P = 0.001) and inducible NO synthase (NOS2) (P = 0.002), suppressed MMP9 (P < 0.001), and impaired phagocytic activity (P < 0.001) upon lipopolysaccharide stimulation. RNA-sequencing revealed grossly altered transcriptomes of MALAT1-deficient splenocytes already at baseline, with massive induction of IFN- γ, TNF, NOS2, and granzyme B; CC and CXC chemokines and CCR8; and innate immunity genes interferon-induced protein with tetratricopeptide repeats (IFIT)1/3, interferon-induced transmembrane protein (IFITM)1/3, ISG15. Multiple miRs were up to 45-fold upregulated. Further, selective ablation of the cytosolic part of the MALAT1 system only, the enzymatically MALAT1-derived mascRNA, resulted in massive induction of TNF (P = 0.004) and IL6 (P = 0.028) in macrophages. Northern analysis of post-myocardial infarction patient vs. control peripheral blood mononuclear cells showed reduced (P = 0.005) mascRNA in the patients. CHART-enriched RNA-sequencing reads at the genomic loci of MALAT1 and neighbouring nuclear enriched abundant transcript (NEAT1) documented direct interaction between these lncRNA transcripts. The data suggest a molecular circuit involving the MALAT1-mascRNA system, interactions between MALAT1 and NEAT1, and key immune effector molecules, cumulatively impacting upon the development of atherosclerosis. It appears reasonable to look for therapeutic targets in this circuit and to screen for anomalies in the NEAT1-MALAT1 region in humans, too, as possible novel disease risk factors.

We present four models of solution free-energy prediction for druglike molecules utilizing cheminformatics descriptors and theoretically calculated thermodynamic values. We make predictions of solution free energy using physics-based theory alone and using machine learning/quantitative structure–property relationship (QSPR) models. We also develop machine learning models where the theoretical energies and cheminformatics descriptors are used as combined input. These models are used to predict solvation free energy. While direct theoretical calculation does not give accurate results in this approach, machine learning is able to give predictions with a root mean squared error (RMSE) of ∼1.1 log S units in a 10-fold cross-validation for our Drug-Like-Solubility-100 (DLS-100) dataset of 100 druglike molecules. We find that a model built using energy terms from our theoretical methodology as descriptors is marginally less predictive than one built on Chemistry Development Kit (CDK) descriptors. Combining both sets of descriptors allows a further but very modest improvement in the predictions. However, in some cases, this is a statistically significant enhancement. These results suggest that there is little complementarity between the chemical information provided by these two sets of descriptors, despite their different sources and methods of calculation. Our machine learning models are also able to predict the well-known Solubility Challenge dataset with an RMSE value of 0.9–1.0 log S units.

Objective : Electrocardiography is the most common tool to diagnose cardiovascular diseases. Annotation, segmentation and rhythm classification of ECGs are challenging tasks, especially in the presence of atrial fibrillation and other arrhythmias. Our aim is to increase the accuracy of heart rhythm estimation by the use of extreme gradient boosting trees and the development of a deep convolutional neural network for ECG segmentation. Approach : We trained a convolutional neural network with waveforms from PhysioNet databases to annotate QRS complexes, P waves, T waves, noise and interbeat ECG segments that characterize the essences of normal and irregular heart beats. We evaluated true positive rates, positive predictive values and mean absolute differences of our annotation based on reference annotations of the QT and MIT-BIH P-wave database. Moreover, we compared the results with standard QRS detectors and Ecgpuwave. Extreme gradient boosting trees were used to determine the heart rhythm based on hand-crafted features. More precisely, a noise estimation function was used in combination with heart rate and interval data. Furthermore we defined particular features based on ECG morphology, appearance of P waves and detection of irregular beats. We examined the feature importance and identified key features for normal sinus rhythm, atrial fibrillation, alternative rhythm and noisy recordings. The classification performance was evaluated externally using F ₁ scores by applying the algorithm to the hidden test set provided by the PhysioNet/CinC Challenge 2017. Main results : The true positive rate of the convolutional neural network in detection of manually revised R peaks in the QT database was and the positive predictive value was . The detection of P and T waves reached a true positive rate of and respectively, given a 50 ms tolerance when comparing the reference to the test annotation set. The rhythm classification performance reached an overall F ₁ score of 0.82 when applying the algorithm to the hidden test set. Significance : We achieved a shared rank #9 in the post-challenge phase of the PhysioNet/CinC Challenge 2017.

Podocyte loss and changes to the complex morphology are major causes of chronic kidney disease (CKD). As the incidence is continuously increasing over the last decades without sufficient treatment, it is important to find predicting biomarkers. Therefore, we measured urinary mRNA levels of podocyte genes NPHS1, NPHS2, PODXL and BDNF, KIM‐1, CTSL by qRT‐PCR of 120 CKD patients. We showed a strong correlation between BDNF and the kidney injury marker KIM‐1, which were also correlated with NPHS1, suggesting podocytes as a contributing source. In human biopsies, BDNF was localized in the cell body and major processes of podocytes. In glomeruli of diabetic nephropathy patients, we found a strong BDNF signal in the remaining podocytes. An inhibition of the BDNF receptor TrkB resulted in enhanced podocyte dedifferentiation. The knockdown of the orthologue resulted in pericardial oedema formation and lowered viability of zebrafish larvae. We found an enlarged Bowman's space, dilated glomerular capillaries, podocyte loss and an impaired glomerular filtration. We demonstrated that BDNF is essential for glomerular development, morphology and function and the expression of BDNF and KIM‐1 is highly correlated in urine cells of CKD patients. Therefore, BDNF mRNA in urine cells could serve as a potential CKD biomarker.