Post-translational modifications (PTMs) are closely linked to numerous diseases, playing a significant role in regulating protein structures, activities, and functions. Therefore, the identification of PTMs is crucial for understanding the mechanisms of cell biology and diseases therapy. Compared to traditional machine learning methods, the deep learning approaches for PTM prediction provide accurate and rapid screening, guiding the downstream wet experiments to leverage the screened information for focused studies. In this paper, we reviewed the recent works in deep learning to identify phosphorylation, acetylation, ubiquitination, and other PTM types. In addition, we summarized PTM databases and discussed future directions with critical insights.

Optimization of the fermentation process for recombinant protein production (RPP) is often resource-intensive. Machine learning (ML) approaches are helpful in minimizing the experimentations and find vast applications in RPP. However, these ML-based tools primarily focus on features with respect to amino-acid-sequence, ruling out the influence of fermentation process conditions. The present study combines the features derived from fermentation process conditions with that from amino acid-sequence to construct an ML-based model that predicts the maximal protein yields and the corresponding fermentation conditions for the expression of target recombinant protein in the Escherichia coli periplasm. Two sets of XGBoost classifiers were employed in the first stage to classify the expression levels of the target protein as high (>50 mg/L), medium (between 0.5 and 50 mg/L), or low (<0.5 mg/L). The second-stage framework consisted of three regression models involving support vector machines and random forest to predict the expression yields corresponding to each expression-level-class. Independent tests showed that the predictor achieved an overall average accuracy of 75% and a Pearson coefficient correlation of 0.91 for the correctly classified instances. Therefore, our model offers a reliable substitution of numerous trial-and-error experiments to identify the optimal fermentation conditions and yield for RPP. It is also implemented as an open-access webserver, PERISCOPE-Opt (http://periscope-opt.erc.monash.edu).

UNASSIGNED: Tumor delineation is required both for radiotherapy planning and quantitative imaging biomarker purposes. It is a manual, time- and labor-intensive process prone to inter- and intraobserver variations. Semi or fully automatic segmentation could provide better efficiency and consistency. This study aimed to investigate the influence of including and combining functional with anatomical magnetic resonance imaging (MRI) sequences on the quality of automatic segmentations.
UNASSIGNED: T2-weighted (T2w), diffusion weighted, multi-echo T2*-weighted, and contrast enhanced dynamic multi-echo (DME) MR images of eighty-one patients with rectal cancer were used in the analysis. Four classical machine learning algorithms; adaptive boosting (ADA), linear and quadratic discriminant analysis and support vector machines, were trained for automatic segmentation of tumor and normal tissue using different combinations of the MR images as input, followed by semi-automatic morphological post-processing. Manual delineations from two experts served as ground truth. The Sørensen-Dice similarity coefficient (DICE) and mean symmetric surface distance (MSD) were used as performance metric in leave-one-out cross validation.
UNASSIGNED: Using T2w images alone, ADA outperformed the other algorithms, yielding a median per patient DICE of 0.67 and MSD of 3.6 mm. The performance improved when functional images were added and was highest for models based on either T2w and DME images (DICE: 0.72, MSD: 2.7 mm) or all four MRI sequences (DICE: 0.72, MSD: 2.5 mm).
UNASSIGNED: Machine learning models using functional MRI, in particular DME, have the potential to improve automatic segmentation of rectal cancer relative to models using T2w MRI alone.

Metabolomics is an expanding field of medical diagnostics since many diseases cause metabolic reprogramming alteration. Additionally, the metabolic point of view offers an insight into the molecular mechanisms of diseases. Due to the complexity of metabolic assignment dependent on the 1D NMR spectral analysis, 2D NMR techniques are preferred because of spectral resolution issues. Thus, in this work, we introduce an automated metabolite identification and assignment from 1H-1H TOCSY (total correlation spectroscopy) using real breast cancer tissue. The new approach is based on customized and extended semi-supervised classifiers: KNFST, SVM, third (PC3) and fourth (PC4) degree polynomial. In our approach, metabolic assignment is based only on the vertical and horizontal frequencies of the metabolites in the 1H-1H TOCSY. KNFST and SVM show high performance (high accuracy and low mislabeling rate) in relatively low size of initially labeled training data. PC3 and PC4 classifiers showed lower accuracy and high mislabeling rates, and both classifiers fail to provide an acceptable accuracy at extremely low size (≤9% of the entire dataset) of initial training data. Additionally, semi-supervised classifiers were implemented to obtain a fully automatic procedure for signal assignment and deconvolution of TOCSY, which is a big step forward in NMR metabolic profiling. A set of 27 metabolites were deduced from the TOCSY, and their assignments agreed with the metabolites deduced from a 1D NMR spectrum of the same sample analyzed by conventional human-based methodology.

UNASSIGNED: Gliomas are one of the most common types of primary tumors in central nervous system. Previous studies have found that macrophages actively participate in tumor growth.
UNASSIGNED: Weighted gene co-expression network analysis was used to identify meaningful macrophage-related gene genes for clustering. Pamr, SVM, and neural network were applied for validating clustering results. Somatic mutation and methylation were used for defining the features of identified clusters. Differentially expressed genes (DEGs) between the stratified groups after performing elastic regression and principal component analyses were used for the construction of MScores. The expression of macrophage-specific genes were evaluated in tumor microenvironment based on single cell sequencing analysis. A total of 2365 samples from 15 glioma datasets and 5842 pan-cancer samples were used for external validation of MScore.
UNASSIGNED: Macrophages were identified to be negatively associated with the survival of glioma patients. Twenty-six macrophage-specific DEGs obtained by elastic regression and PCA were highly expressed in macrophages at single-cell level. The prognostic value of MScores in glioma was validated by the active proinflammatory and metabolic profile of infiltrating microenvironment and response to immunotherapies of samples with this signature. MScores managed to stratify patient survival probabilities in 15 external glioma datasets and pan-cancer datasets, which predicted worse survival outcome. Sequencing data and immunohistochemistry of Xiangya glioma cohort confirmed the prognostic value of MScores. A prognostic model based on MScores demonstrated high accuracy rate.
UNASSIGNED: Our findings strongly support a modulatory role of macrophages, especially M2 macrophages in glioma progression and warrants further experimental studies.

Drug discovery aims at finding new compounds with specific chemical properties for the treatment of diseases. In the last years, the approach used in this search presents an important component in computer science with the skyrocketing of machine learning techniques due to its democratization. With the objectives set by the Precision Medicine initiative and the new challenges generated, it is necessary to establish robust, standard and reproducible computational methodologies to achieve the objectives set. Currently, predictive models based on Machine Learning have gained great importance in the step prior to preclinical studies. This stage manages to drastically reduce costs and research times in the discovery of new drugs. This review article focuses on how these new methodologies are being used in recent years of research. Analyzing the state of the art in this field will give us an idea of where cheminformatics will be developed in the short term, the limitations it presents and the positive results it has achieved. This review will focus mainly on the methods used to model the molecular data, as well as the biological problems addressed and the Machine Learning algorithms used for drug discovery in recent years.

Advances in nucleic acid sequencing technology have enabled expansion of our ability to profile microbial diversity. These large datasets of taxonomic and functional diversity are key to better understanding microbial ecology. Machine learning has proven to be a useful approach for analyzing microbial community data and making predictions about outcomes including human and environmental health. Machine learning applied to microbial community profiles has been used to predict disease states in human health, environmental quality and presence of contamination in the environment, and as trace evidence in forensics. Machine learning has appeal as a powerful tool that can provide deep insights into microbial communities and identify patterns in microbial community data. However, often machine learning models can be used as black boxes to predict a specific outcome, with little understanding of how the models arrived at predictions. Complex machine learning algorithms often may value higher accuracy and performance at the sacrifice of interpretability. In order to leverage machine learning into more translational research related to the microbiome and strengthen our ability to extract meaningful biological information, it is important for models to be interpretable. Here we review current trends in machine learning applications in microbial ecology as well as some of the important challenges and opportunities for more broad application of machine learning to understanding microbial communities.

(123)I-ioflupane single photon emission computed tomography (SPECT) is a sensitive and well established imaging tool in Parkinson\'s disease (PD) and atypical parkinsonian syndromes (APS), yet a discrimination between PD and APS has been considered inconsistent at least based on visual inspection or simple region of interest analyses. We here reappraise this issue by applying advanced image analysis techniques to separate PD from the various APS. This study included 392 consecutive patients with degenerative parkinsonism undergoing (123)I-ioflupane SPECT at our institution over the last decade: 306 PD, 24 multiple system atrophy (MSA), 32 progressive supranuclear palsy (PSP) and 30 corticobasal degeneration (CBD) patients. Data analysis included voxel-wise univariate statistical parametric mapping and multivariate pattern recognition using linear discriminant classifiers. MSA and PSP showed less ioflupane uptake in the head of caudate nucleus relative to PD and CBD, yet there was no difference between MSA and PSP. CBD had higher uptake in both putamen relative to PD, MSA and PSP. Classification was significant for PD versus APS (AUC 0.69, p < 0.05) and between APS subtypes (MSA vs CBD AUC 0.80, p < 0.05; MSA vs PSP AUC 0.69 p < 0.05; CBD vs PSP AUC 0.69 p < 0.05). Both striatal and extra-striatal regions contain classification information, yet the combination of both regions does not significantly improve classification accuracy. PD, MSA, PSP and CBD have distinct patterns of dopaminergic depletion on (123)I-ioflupane SPECT. The high specificity of 84-90% for PD versus APS indicates that the classifier is particularly useful for confirming APS cases.