Heuristic-enabled active machine learning: A case study of predicting essential developmental stage and immune response genes in Drosophila melanogaster.

Computational prediction of absolute essential genes using machine learning has gained wide attention in recent years. However, essential genes are mostly conditional and not absolute. Experimental techniques provide a reliable approach of identifying conditionally essential genes; however, experimental methods are laborious, time and resource consuming, hence computational techniques have been used to complement the experimental methods. Computational techniques such as supervised machine learning, or flux balance analysis are grossly limited due to the unavailability of required data for training the model or simulating the conditions for gene essentiality. This study developed a heuristic-enabled active machine learning method based on a light gradient boosting model to predict essential immune response and embryonic developmental genes in Drosophila melanogaster. We proposed a new sampling selection technique and introduced a heuristic function which replaces the human component in traditional active learning models. The heuristic function dynamically selects the unlabelled samples to improve the performance of the classifier in the next iteration. Testing the proposed model with four benchmark datasets, the proposed model showed superior performance when compared to traditional active learning models (random sampling and uncertainty sampling). Applying the model to identify conditionally essential genes, four novel essential immune response genes and a list of 48 novel genes that are essential in embryonic developmental condition were identified. We performed functional enrichment analysis of the predicted genes to elucidate their biological processes and the result evidence our predictions. Immune response and embryonic development related processes were significantly enriched in the essential immune response and embryonic developmental genes, respectively. Finally, we propose the predicted essential genes for future experimental studies and use of the developed tool accessible at http://heal.covenantuniversity.edu.ng for conditional essentiality predictions.

Machine learning on large scale perturbation screens for SARS-CoV-2 host factors identifies ß-catenin/CBP inhibitor PRI-724 as a potent antiviral.

Expanding antiviral treatment options against SARS-CoV-2 remains crucial as the virus evolves under selection pressure which already led to the emergence of several drug resistant strains. Broad spectrum host-directed antivirals (HDA) are promising therapeutic options, however the robust identification of relevant host factors by CRISPR/Cas9 or RNA interference screens remains challenging due to low consistency in the resulting hits. To address this issue, we employed machine learning, based on experimental data from several knockout screens and a drug screen. We trained classifiers using genes essential for virus life cycle obtained from the knockout screens. The machines based their predictions on features describing cellular localization, protein domains, annotated gene sets from Gene Ontology, gene and protein sequences, and experimental data from proteomics, phospho-proteomics, protein interaction and transcriptomic profiles of SARS-CoV-2 infected cells. The models reached a remarkable performance suggesting patterns of intrinsic data consistency. The predicted HDF were enriched in sets of genes particularly encoding development, morphogenesis, and neural processes. Focusing on development and morphogenesis-associated gene sets, we found ß-catenin to be central and selected PRI-724, a canonical ß-catenin/CBP disruptor, as a potential HDA. PRI-724 limited infection with SARS-CoV-2 variants, SARS-CoV-1, MERS-CoV and IAV in different cell line models. We detected a concentration-dependent reduction in cytopathic effects, viral RNA replication, and infectious virus production in SARS-CoV-2 and SARS-CoV-1-infected cells. Independent of virus infection, PRI-724 treatment caused cell cycle deregulation which substantiates its potential as a broad spectrum antiviral. Our proposed machine learning concept supports focusing and accelerating the discovery of host dependency factors and identification of potential host-directed antivirals.

Anopheles gambiae Trehalase Inhibitors for Malaria Vector Control: A Molecular Docking and Molecular Dynamics Study.

Trehalase inhibitors are considered safe alternatives for insecticides and fungicides. However, there are no studies testing these compounds on Anopheles gambiae, a major vector of human malaria. This study predicted the three-dimensional structure of Anopheles gambiae trehalase (AgTre) and identified potential inhibitors using molecular docking and molecular dynamics methods. Robetta server, C-I-TASSER, and I-TASSER were used to predict the protein structure, while the structural assessment was carried out using SWISS-MODEL, ERRAT, and VERIFY3D. Molecular docking and screening of 3022 compounds was carried out using AutoDock Vina in PyRx, and MD simulation was carried out using NAMD. The Robetta model outperformed all other models and was used for docking and simulation studies. After a post-screening analysis and ADMET studies, uniflorine, 67837201, 10406567, and Compound 2 were considered the best hits with binding energies of -6.9, -8.9, -9, and -8.4 kcal/mol, respectively, better than validamycin A standard (-5.4 kcal/mol). These four compounds were predicted to have no eco-toxicity, Brenk, or PAINS alerts. Similarly, they were predicted to be non-mutagenic, carcinogenic, or hepatoxic. 67837201, 10406567, and Compound 2 showed excellent stability during simulation. The study highlights uniflorine, 67837201, 10406567, and Compound 2 as good inhibitors of AgTre and possible compounds for malaria vector control.

Intrafamilial variability in six family members with ERF-related craniosynostosis syndrome type 4.

ERF-related craniosynostosis syndrome type 4 (CRS4, OMIM #600775) is a rare autosomal dominant malformation syndrome, caused by pathogenic variants in the ERF gene and characterized by craniosynostosis, developmental delay, and dysmorphic features such as hypertelorism, exophthalmos, depressed nasal bridge, and retrognathia. So far, there are mostly individual reports and only a few descriptions of families with more than two affected patients, allowing statements about the penetrance of a certain variant and its variability only to a limited extent. In this study, we report an in-depth analysis of the clinical course of six family members from three generations with the novel heterozygous nonsense variant c.286A>T (p.Lys96*) in the ERF gene. At the time of examination, all of the six patients showed mild dysmorphic features and brachydactyly, five were overweight/obese and had delayed speech development, and four were short in stature. Hyperactivity, a short concentration span and a history of learning difficulties were found in half of the affected family members. To this day, none of the patients developed increased intracranial hypertension that would require surgical intervention. This work provides further information on the expressive variability of an ERF variant in six members of one family and focuses on the need for close neuropediatric surveillance.

Linear programming based gene expression model (LPM-GEM) predicts the carbon source for Bacillus subtilis.

BACKGROUND: Elucidating cellular metabolism led to many breakthroughs in biotechnology, synthetic biology, and health sciences. To date, deriving metabolic fluxes by 13C tracer experiments is the most prominent approach for studying metabolic fluxes quantitatively, often with high accuracy and precision. However, the technique has a high demand for experimental resources. Alternatively, flux balance analysis (FBA) has been employed to estimate metabolic fluxes without labeling experiments. It is less informative but can benefit from the low costs and low experimental efforts and gain flux estimates in experimentally difficult conditions. Methods to integrate relevant experimental data have been emerged to improve FBA flux estimations. Data from transcription profiling is often selected since it is easy to generate at the genome scale, typically embedded by a discretization of differential and non-differential expressed genes coding for the respective enzymes. RESULT: We established the novel method Linear Programming based Gene Expression Model (LPM-GEM). LPM-GEM linearly embeds gene expression into FBA constraints. We implemented three strategies to reduce thermodynamically infeasible loops, which is a necessary prerequisite for such an omics-based model building. As a case study, we built a model of B. subtilis grown in eight different carbon sources. We obtained good flux predictions based on the respective transcription profiles when validating with 13C tracer based metabolic flux data of the same conditions. We could well predict the specific carbon sources. When testing the model on another, unseen dataset that was not used during training, good prediction performance was also observed. Furthermore, LPM-GEM outperformed a well-established model building methods. CONCLUSION: Employing LPM-GEM integrates gene expression data efficiently. The method supports gene expression-based FBA models and can be applied as an alternative to estimate metabolic fluxes when tracer experiments are inappropriate.

PITX1 Is a Regulator of TERT Expression in Prostate Cancer with Prognostic Power.

The current risk stratification in prostate cancer (PCa) is frequently insufficient to adequately predict disease development and outcome. One hallmark of cancer is telomere maintenance. For telomere maintenance, PCa cells exclusively employ telomerase, making it essential for this cancer entity. However, TERT, the catalytic protein component of the reverse transcriptase telomerase, itself does not suit as a prognostic marker for prostate cancer as it is rather low expressed. We investigated if, instead of TERT, transcription factors regulating TERT may suit as prognostic markers. To identify transcription factors regulating TERT, we developed and applied a new gene regulatory modeling strategy to a comprehensive transcriptome dataset of 445 primary PCa. Six transcription factors were predicted as TERT regulators, and most prominently, the developmental morphogenic factor PITX1. PITX1 expression positively correlated with telomere staining intensity in PCa tumor samples. Functional assays and chromatin immune-precipitation showed that PITX1 activates TERT expression in PCa cells. Clinically, we observed that PITX1 is an excellent prognostic marker, as concluded from an analysis of more than 15,000 PCa samples. PITX1 expression in tumor samples associated with (i) increased Ki67 expression indicating increased tumor growth, (ii) a worse prognosis, and (iii) correlated with telomere length.

Computational identification of Plasmodium falciparum RNA pseudouridylate synthase as a viable drug target, its physicochemical properties, 3D structure prediction and prediction of potential inhibitors.

The increased resistance to the currently effective antimalarial drugs against Plasmodium falciparum has necessitated the development of new drugs for malaria treatment. Many proteins have been predicted using various means as potential drug targets for the treatment of the P. falciparum malaria infection. Meanwhile, only a few studies went on to predict the 3-dimensional (3D) structure of potential target. Therefore, this study aimed to predict potential antimalarial drug targets against the deadliest malaria parasite P. falciparum as well as to determine the 3D structure and possible inhibitors of one of the targets. We employed machine learning approach to predict suitable drug targets in P. falciparum. Five of the predicted protein targets were considered as potential drug targets as they were non-homologous to their human counterparts. Out of these, we determined the physicochemical properties, predicted the 3D structure and carried out docking-based virtual screening of P. falciparum RNA pseudouridylate synthase, putative (PfRPuSP). The PfRPuSP was one of the potential five target proteins. Homology modelling and the ab initio methods were used to predict the 3D structure of PfRPuSP. Then, a compound library of 5621 molecules was constructed from PubChem and ChEMBL databases using 5-fluorouridine as the control inhibitor. Docking-based virtual screening was performed using Autodock 4.2 and Autodock Vina to select compounds with high binding affinity. A total of 11 compounds were selected based on their binding energies from 881 compounds which were manually examined after docking. Seven of the 11 compounds that exhibited remarkable interactions with the residues in the active sites of PfRPuSP were analysed. These compounds performed favourably when compared to the control inhibitor and predicted to bind better than 5-fluorouridine. These seven compounds are suggested as new potential lead structures for antimalarial treatment.

Identifying essential genes across eukaryotes by machine learning.

Identifying essential genes on a genome scale is resource intensive and has been performed for only a few eukaryotes. For less studied organisms essentiality might be predicted by gene homology. However, this approach cannot be applied to non-conserved genes. Additionally, divergent essentiality information is obtained from studying single cells or whole, multi-cellular organisms, and particularly when derived from human cell line screens and human population studies. We employed machine learning across six model eukaryotes and 60 381 genes, using 41 635 features derived from the sequence, gene function information and network topology. Within a leave-one-organism-out cross-validation, the classifiers showed high generalizability with an average accuracy close to 80% in the left-out species. As a case study, we applied the method to Tribolium castaneum and Bombyx mori and validated predictions experimentally yielding similar performances. Finally, using the classifier based on the studied model organisms enabled linking the essentiality information of human cell line screens and population studies.

Predicting host dependency factors of pathogens in Drosophila melanogaster using machine learning.

Pathogens causing infections, and particularly when invading the host cells, require the host cell machinery for efficient regeneration and proliferation during infection. For their life cycle, host proteins are needed and these Host Dependency Factors (HDF) may serve as therapeutic targets. Several attempts have approached screening for HDF producing large lists of potential HDF with, however, only marginal overlap. To get consistency into the data of these experimental studies, we developed a machine learning pipeline. As a case study, we used publicly available lists of experimentally derived HDF from twelve different screening studies based on gene perturbation in Drosophila melanogaster cells or in vivo upon bacterial or protozoan infection. A total of 50,334 gene features were generated from diverse categories including their functional annotations, topology attributes in protein interaction networks, nucleotide and protein sequence features, homology properties and subcellular localization. Cross-validation revealed an excellent prediction performance. All feature categories contributed to the model. Predicted and experimentally derived HDF showed a good consistency when investigating their common cellular processes and function. Cellular processes and molecular function of these genes were highly enriched in membrane trafficking, particularly in the trans-Golgi network, cell cycle and the Rab GTPase binding family. Using our machine learning approach, we show that HDF in organisms can be predicted with high accuracy evidencing their common investigated characteristics. We elucidated cellular processes which are utilized by invading pathogens during infection. Finally, we provide a list of 208 novel HDF proposed for future experimental studies.

Anopheles metabolic proteins in malaria transmission, prevention and control: a review.

The increasing resistance to currently available insecticides in the malaria vector, Anopheles mosquitoes, hampers their use as an effective vector control strategy for the prevention of malaria transmission. Therefore, there is need for new insecticides and/or alternative vector control strategies, the development of which relies on the identification of possible targets in Anopheles. Some known and promising targets for the prevention or control of malaria transmission exist among Anopheles metabolic proteins. This review aims to elucidate the current and potential contribution of Anopheles metabolic proteins to malaria transmission and control. Highlighted are the roles of metabolic proteins as insecticide targets, in blood digestion and immune response as well as their contribution to insecticide resistance and Plasmodium parasite development. Furthermore, strategies by which these metabolic proteins can be utilized for vector control are described. Inhibitors of Anopheles metabolic proteins that are designed based on target specificity can yield insecticides with no significant toxicity to non-target species. These metabolic modulators combined with each other or with synergists, sterilants, and transmission-blocking agents in a single product, can yield potent malaria intervention strategies. These combinations can provide multiple means of controlling the vector. Also, they can help to slow down the development of insecticide resistance. Moreover, some metabolic proteins can be modulated for mosquito population replacement or suppression strategies, which will significantly help to curb malaria transmission.

Essential gene prediction in Drosophila melanogaster using machine learning approaches based on sequence and functional features.

Genes are termed to be essential if their loss of function compromises viability or results in profound loss of fitness. On the genome scale, these genes can be determined experimentally employing RNAi or knockout screens, but this is very resource intensive. Computational methods for essential gene prediction can overcome this drawback, particularly when intrinsic (e.g. from the protein sequence) as well as extrinsic features (e.g. from transcription profiles) are considered. In this work, we employed machine learning to predict essential genes in Drosophila melanogaster. A total of 27,340 features were generated based on a large variety of different aspects comprising nucleotide and protein sequences, gene networks, protein-protein interactions, evolutionary conservation and functional annotations. Employing cross-validation, we obtained an excellent prediction performance. The best model achieved in D. melanogaster a ROC-AUC of 0.90, a PR-AUC of 0.30 and a F1 score of 0.34. Our approach considerably outperformed a benchmark method in which only features derived from the protein sequences were used (P < 0.001). Investigating which features contributed to this success, we found all categories of features, most prominently network topological, functional and sequence-based features. To evaluate our approach we performed the same workflow for essential gene prediction in human and achieved an ROC-AUC = 0.97, PR-AUC = 0.73, and F1 = 0.64. In summary, this study shows that using our well-elaborated assembly of features covering a broad range of intrinsic and extrinsic gene and protein features enabled intelligent systems to predict well the essentiality of genes in an organism.

Biallelic variants in KYNU cause a multisystemic syndrome with hand hyperphalangism.

Catel-Manzke syndrome is characterized by the combination of Pierre Robin sequence and radial deviation, shortening as well as clinodactyly of the index fingers, due to an accessory ossification center. Mutations in TGDS have been identified as one cause of Catel-Manzke syndrome, but cannot be found as causative in every patient with the clinical diagnosis. We performed a chromosome microarray and/or exome sequencing in three patients with hand hyperphalangism, heart defect, short stature, and mild to severe developmental delay, all of whom were initially given a clinical diagnosis of Catel-Manzke syndrome. In one patient, we detected a large deletion of exons 1-8 and the missense variant c.1282C > T (p.Arg428Trp) in KYNU (NM_003937.2), whereas homozygous missense variants in KYNU were found in the other two patients (c.989G > A (p.Arg330Gln) and c.326G > C (p.Trp109Ser)). Plasma and urine metabolomic analysis of two patients indicated a block along the tryptophan catabolic pathway and urine organic acid analysis showed excretion of xanthurenic acid. Biallelic loss-of-function mutations in KYNU were recently described as a cause of NAD deficiency with vertebral, cardiac, renal and limb defects; however, no hand hyperphalangism was described in those patients, and Catel-Manzke syndrome was not discussed as a differential diagnosis. In conclusion, we present unrelated patients identified with biallelic variants in KYNU leading to kynureninase deficiency and xanthurenic aciduria as a very likely cause of their hyperphalangism, heart defect, short stature, and developmental delay. We suggest performance of urine organic acid analysis in patients with suspected Catel-Manzke syndrome, particularly in those with cardiac or vertebral defects or without mutations in TGDS.

SOPH syndrome in three affected individuals showing similarities with progeroid cutis laxa conditions in early infancy.

Individuals affected with autosomal recessive cutis laxa type 2B and 3 usually show translucent skin with visible veins and abnormal elastic fibers, intrauterine and/or postnatal growth restriction and a typical triangular facial gestalt. Here we describe three unrelated individuals in whom such a cutis laxa syndrome was suspected, especially after electron microscopy revealed immature and less dense dermal elastic fibers in one of them. However, one of these children also displayed optic atrophy and two hypogammaglobulinemia. All had elevated liver enzymes and acute liver failure during febrile episodes leading to early demise in two of them. The only surviving patient had been treated with immunoglobulins. Through exome sequencing we identified mutations in NBAS, coding for a protein involved in Golgi-to-ER transport. NBAS deficiency causes several rare conditions ranging from isolated recurrent acute liver failure to a multisystem disorder mainly characterized by short stature, optic nerve atrophy and Pelger-Huët anomaly (SOPH). Since we subsequently verified Pelger-Huët anomaly in two of the patients the diagnosis SOPH syndrome was unequivocally proven. Our data show that SOPH syndrome can be regarded as a differential diagnosis for the progeroid forms of cutis laxa in early infancy and that possibly treatment of the hypogammaglobulinemia can be of high relevance for the prognosis.

Correction: Loss of the BMP Antagonist, SMOC-1, Causes Ophthalmo-Acromelic (Waardenburg Anophthalmia) Syndrome in Humans and Mice.

[This corrects the article DOI: 10.1371/journal.pgen.1002114.].

De Novo Mutations in SLC25A24 Cause a Craniosynostosis Syndrome with Hypertrichosis, Progeroid Appearance, and Mitochondrial Dysfunction.

Gorlin-Chaudhry-Moss syndrome (GCMS) is a dysmorphic syndrome characterized by coronal craniosynostosis and severe midface hypoplasia, body and facial hypertrichosis, microphthalmia, short stature, and short distal phalanges. Variable lipoatrophy and cutis laxa are the basis for a progeroid appearance. Using exome and genome sequencing, we identified the recurrent de novo mutations c.650G>A (p.Arg217His) and c.649C>T (p.Arg217Cys) in SLC25A24 in five unrelated girls diagnosed with GCMS. Two of the girls had pronounced neonatal progeroid features and were initially diagnosed with Wiedemann-Rautenstrauch syndrome. SLC25A24 encodes a mitochondrial inner membrane ATP-Mg/Pi carrier. In fibroblasts from affected individuals, the mutated SLC25A24 showed normal stability. In contrast to control cells, the probands' cells showed mitochondrial swelling, which was exacerbated upon treatment with hydrogen peroxide (H2O2). The same effect was observed after overexpression of the mutant cDNA. Under normal culture conditions, the mitochondrial membrane potential of the probands' fibroblasts was intact, whereas ATP content in the mitochondrial matrix was lower than that in control cells. However, upon H2O2 exposure, the membrane potential was significantly elevated in cells harboring the mutated SLC25A24. No reduction of mitochondrial DNA copy number was observed. These findings demonstrate that mitochondrial dysfunction with increased sensitivity to oxidative stress is due to the SLC25A24 mutations. Our results suggest that the SLC25A24 mutations induce a gain of pathological function and link mitochondrial ATP-Mg/Pi transport to the development of skeletal and connective tissue.

Homozygous and compound-heterozygous mutations in TGDS cause Catel-Manzke syndrome.

Catel-Manzke syndrome is characterized by Pierre Robin sequence and a unique form of bilateral hyperphalangy causing a clinodactyly of the index finger. We describe the identification of homozygous and compound heterozygous mutations in TGDS in seven unrelated individuals with typical Catel-Manzke syndrome by exome sequencing. Six different TGDS mutations were detected: c.892A>G (p.Asn298Asp), c.270_271del (p.Lys91Asnfs(∗)22), c.298G>T (p.Ala100Ser), c.294T>G (p.Phe98Leu), c.269A>G (p.Glu90Gly), and c.700T>C (p.Tyr234His), all predicted to be disease causing. By using haplotype reconstruction we showed that the mutation c.298G>T is probably a founder mutation. Due to the spectrum of the amino acid changes, we suggest that loss of function in TGDS is the underlying mechanism of Catel-Manzke syndrome. TGDS (dTDP-D-glucose 4,6-dehydrogenase) is a conserved protein belonging to the SDR family and probably plays a role in nucleotide sugar metabolism.

Deletions in the 3' part of the NFIX gene including a recurrent Alu-mediated deletion of exon 6 and 7 account for previously unexplained cases of Marshall-Smith syndrome.

Marshall-Smith syndrome (MSS) is a very rare malformation syndrome characterized by typical craniofacial anomalies, abnormal osseous maturation, developmental delay, failure to thrive, and respiratory difficulties. Mutations in the nuclear factor 1/X gene (NFIX) were recently identified as the cause of MSS. In our study cohort of 17 patients with a clinical diagnosis of MSS, conventional sequencing of NFIX revealed frameshift and splice-site mutations in 10 individuals. Using multiplex ligation-dependent probe amplification analysis, we identified a recurrent deletion of NFIX exon 6 and 7 in five individuals. We demonstrate this recurrent deletion is the product of a recombination between AluY elements located in intron 5 and 7. Two other patients had smaller deletions affecting exon 6. These findings show that MSS is a genetically homogeneous Mendelian disorder. RT-PCR experiments with newly identified NFIX mutations including the recurrent exon 6 and 7 deletion confirmed previous findings indicating that MSS-associated mutant mRNAs are not cleared by nonsense-mediated mRNA decay. Predicted MSS-associated mutant NFIX proteins consistently have a preserved DNA binding and dimerization domain, whereas they grossly vary in their C-terminal portion. This is in line with the hypothesis that MSS-associated mutations encode dysfunctional proteins that act in a dominant negative manner.

Loss of the BMP antagonist, SMOC-1, causes Ophthalmo-acromelic (Waardenburg Anophthalmia) syndrome in humans and mice.

Ophthalmo-acromelic syndrome (OAS), also known as Waardenburg Anophthalmia syndrome, is defined by the combination of eye malformations, most commonly bilateral anophthalmia, with post-axial oligosyndactyly. Homozygosity mapping and subsequent targeted mutation analysis of a locus on 14q24.2 identified homozygous mutations in SMOC1 (SPARC-related modular calcium binding 1) in eight unrelated families. Four of these mutations are nonsense, two frame-shift, and two missense. The missense mutations are both in the second Thyroglobulin Type-1 (Tg1) domain of the protein. The orthologous gene in the mouse, Smoc1, shows site- and stage-specific expression during eye, limb, craniofacial, and somite development. We also report a targeted pre-conditional gene-trap mutation of Smoc1 (Smoc1(tm1a)) that reduces mRNA to ∼10% of wild-type levels. This gene-trap results in highly penetrant hindlimb post-axial oligosyndactyly in homozygous mutant animals (Smoc1(tm1a/tm1a)). Eye malformations, most commonly coloboma, and cleft palate occur in a significant proportion of Smoc1(tm1a/tm1a) embryos and pups. Thus partial loss of Smoc-1 results in a convincing phenocopy of the human disease. SMOC-1 is one of the two mammalian paralogs of Drosophila Pentagone, an inhibitor of decapentaplegic. The orthologous gene in Xenopus laevis, Smoc-1, also functions as a Bone Morphogenic Protein (BMP) antagonist in early embryogenesis. Loss of BMP antagonism during mammalian development provides a plausible explanation for both the limb and eye phenotype in humans and mice.