Benchmarking antibody clustering methods using sequence, structural, and machine learning similarity measures for antibody discovery applications.

Antibodies are proteins produced by our immune system that have been harnessed as biotherapeutics. The discovery of antibody-based therapeutics relies on analyzing large volumes of diverse sequences coming from phage display or animal immunizations. Identification of suitable therapeutic candidates is achieved by grouping the sequences by their similarity and subsequent selection of a diverse set of antibodies for further tests. Such groupings are typically created using sequence-similarity measures alone. Maximizing diversity in selected candidates is crucial to reducing the number of tests of molecules with near-identical properties. With the advances in structural modeling and machine learning, antibodies can now be grouped across other diversity dimensions, such as predicted paratopes or three-dimensional structures. Here we benchmarked antibody grouping methods using clonotype, sequence, paratope prediction, structure prediction, and embedding information. The results were benchmarked on two tasks: binder detection and epitope mapping. We demonstrate that on binder detection no method appears to outperform the others, while on epitope mapping, clonotype, paratope, and embedding clusterings are top performers. Most importantly, all the methods propose orthogonal groupings, offering more diverse pools of candidates when using multiple methods than any single method alone. To facilitate exploring the diversity of antibodies using different methods, we have created an online tool-CLAP-available at (clap.naturalantibody.com) that allows users to group, contrast, and visualize antibodies using the different grouping methods.

LAP: Liability Antibody Profiler by sequence & structural mapping of natural and therapeutic antibodies.

Antibody-based therapeutics must not undergo chemical modifications that would impair their efficacy or hinder their developability. A commonly used technique to de-risk lead biotherapeutic candidates annotates chemical liability motifs on their sequence. By analyzing sequences from all major sources of data (therapeutics, patents, GenBank, literature, and next-generation sequencing outputs), we find that almost all antibodies contain an average of 3-4 such liability motifs in their paratopes, irrespective of the source dataset. This is in line with the common wisdom that liability motif annotation is over-predictive. Therefore, we have compiled three computational flags to prioritize liability motifs for removal from lead drug candidates: 1. germline, to reflect naturally occurring motifs, 2. therapeutic, reflecting chemical liability motifs found in therapeutic antibodies, and 3. surface, indicative of structural accessibility for chemical modification. We show that these flags annotate approximately 60% of liability motifs as benign, that is, the flagged liabilities have a smaller probability of undergoing degradation as benchmarked on two experimental datasets covering deamidation, isomerization, and oxidation. We combined the liability detection and flags into a tool called Liability Antibody Profiler (LAP), publicly available at lap.naturalantibody.com. We anticipate that LAP will save time and effort in de-risking therapeutic molecules.

Data integration through canonical correlation analysis and its application to OMICs research.

The subject of the paper is a review of multidimensional data analysis methods, which is the canonical analysis with its various variants and its use in omics data research. The dynamic development of high-throughput methods, and with them the availability of large and constantly growing data resources, forces the development of new analytical approaches that allow the review of the analyzed processes, taking into account data from various levels of the organization of living organisms. The multidimensional perspective allows for the assessment of the analyzed phenomenon in a more realistic way, as it generally takes into account much more data (including OMICs data). Without omitting the complexity of an organism, the method simplifies the multidimensional view, finally giving the result so that the researcher can draw practical conclusions. This is particularly important in medical sciences, where the study of pathological processes is usually aimed at developing treatment regimens. One of the primary methods for studying biomedical processes in a multidimensional approach is the canonical correlation analysis (CCA) with various variants. The use of CCA unique methodologies for simultaneous analysis of multiset biomolecular data opens up new avenues for studying previously undiscovered processes and interdependencies such as e.g. in the tumor microenvironment (TME) connected to intercellular communication. Because of the huge and still untapped potential of canonical correlation, in this review available implementations of CCA techniques are presented. In particular, the possibility of using the technique of canonical correlation analysis for OMICs data is emphasized.

Structural modeling of antibody variable regions using deep learning-progress and perspectives on drug discovery.

AlphaFold2 has hallmarked a generational improvement in protein structure prediction. In particular, advances in antibody structure prediction have provided a highly translatable impact on drug discovery. Though AlphaFold2 laid the groundwork for all proteins, antibody-specific applications require adjustments tailored to these molecules, which has resulted in a handful of deep learning antibody structure predictors. Herein, we review the recent advances in antibody structure prediction and relate them to their role in advancing biologics discovery.

Large extracellular vesicles do not mitigate the harmful effect of hyperglycemia on endothelial cell mobility.

Extracellular vesicles, especially the larger fraction (LEVs - large extracellular vesicles), are believed to be an important means of intercellular communication. Earlier studies on LEVs have shown their healing properties, especially in the vascular cells of diabetic patients. Uptake of LEVs by endothelial cells and internalization of their cargo have also been demonstrated. Endothelial cells change their properties under hyperglycemic conditions (HGC), which reduces their activity and is the cause of endothelial dysfunction. The aim of our study was to investigate how human umbilical vein endothelial cells (HUVECs) change their biological properties: shape, mobility, cell surface stiffness, as well as describe the activation of metabolic pathways after exposure to the harmful effects of HGC and the administration of LEVs released by endothelial cells. We obtained LEVs from HUVEC cultures in HGC and normoglycemia (NGC) using the filtration and ultracentrifugation methods. We assessed the size of LEVs and the presence of biomarkers such as phosphatidylserine, CD63, beta-actin and HSP70. We analyzed the LEVs uptake efficiency by HUVECs, HUVEC shape, actin cytoskeleton remodeling, surface stiffness and finally gene expression by mRNA analysis. Under HGC conditions, HUVECs were larger and had a stiffened surface and a strengthened actin cortex compared to cells under NGC condition. HGC also altered the activation of metabolic pathways, especially those related to intracellular transport, metabolism, and organization of cellular components. The most interesting observation in our study is that LEVs did not restore cell motility disturbed by HGC. Although, LEVs were not able to reverse this deleterious effect of HGC, they activated transcription of genes involved in protein synthesis and vesicle trafficking in HUVECs.

Machine-designed biotherapeutics: opportunities, feasibility and advantages of deep learning in computational antibody discovery.

Antibodies are versatile molecular binders with an established and growing role as therapeutics. Computational approaches to developing and designing these molecules are being increasingly used to complement traditional lab-based processes. Nowadays, in silico methods fill multiple elements of the discovery stage, such as characterizing antibody-antigen interactions and identifying developability liabilities. Recently, computational methods tackling such problems have begun to follow machine learning paradigms, in many cases deep learning specifically. This paradigm shift offers improvements in established areas such as structure or binding prediction and opens up new possibilities such as language-based modeling of antibody repertoires or machine-learning-based generation of novel sequences. In this review, we critically examine the recent developments in (deep) machine learning approaches to therapeutic antibody design with implications for fully computational antibody design.

OmicsON - Integration of omics data with molecular networks and statistical procedures.

A huge amount of atomized biological data collected in various databases and the need for a description of their relation by theoretical methods causes the development of data integration methods. The omics data analysis by integration of biological knowledge with mathematical procedures implemented in the OmicsON R library is presented in the paper. OmicsON is a tool for the integration of two sets of data: transcriptomics and metabolomics. In the workflow of the library, the functional grouping and statistical analysis are applied. Subgroups among the transcriptomic and metabolomics sets are created based on the biological knowledge stored in Reactome and String databases. It gives the possibility to analyze such sets of data by multivariate statistical procedures like Canonical Correlation Analysis (CCA) or Partial Least Squares (PLS). The integration of metabolomic and transcriptomic data based on the methodology contained in OmicsON helps to easily obtain information on the connection of data from two different sets. This information can significantly help in assessing the relationship between gene expression and metabolite concentrations, which in turn facilitates the biological interpretation of the analyzed process.

The Clinical Trial Landscape for Melanoma Therapies.

(1) Despite many years of research, melanoma still remains a big challenge for modern medicine. The purpose of this article is to review publicly available clinical trials to find trends regarding the number of trials, their location, and interventions including the most frequently studied drugs and their combinations. (2) We surveyed clinical trials registered in the International Clinical Trials Registry Platform (ICTRP), one of the largest databases on clinical trials. The search was performed on 30 November 2018 using the term "melanoma". Data have been supplemented with the information obtained from publicly available data repositories including PubMed, World Health Organization, National Cancer Institute, Centers for Disease Control and Prevention, European Cancer Information System, and many others to bring the historical context of this study. (3) Among the total of 2563 clinical trials included in the analysis, most have been registered in the USA (1487), which is 58% of the total. The most commonly studied drug in clinical trials was ipilimumab, described as applied intervention in 251 trials. (4) An increase in the number of melanoma clinical trials using immunomodulating monoclonal antibody therapies, small molecule-targeted therapies (inhibitors of BRAF, MEK, CDK4/6), and combination therapies is recognized. This illustrates the tendency towards precision medicine.

Identification of Differentiating Metabolic Pathways between Infant Gut Microbiome Populations Reveals Depletion of Function-Level Adaptation to Human Milk in the Finnish Population.

A variety of autoimmune and allergy events are becoming increasingly common, especially in Western countries. Some pieces of research link such conditions with the composition of microbiota during infancy. In this period, the predominant form of nutrition for gut microbiota is oligosaccharides from human milk (HMO). A number of gut-colonizing strains, such as Bifidobacterium and Bacteroides, are able to utilize HMO, but only some Bifidobacterium strains have evolved to digest the specific composition of human oligosaccharides. Differences in the proportions of the two genera that are able to utilize HMO have already been associated with the frequency of allergies and autoimmune diseases in the Finnish and the Russian populations. Our results show that differences in terms of the taxonomic annotation do not explain the reason for the differences in the Bifidobacterium/Bacteroides ratio between the Finnish and the Russian populations. In this paper, we present the results of function-level analysis. Unlike the typical workflow for gene abundance analysis, BiomeScout technology explains the differences in the Bifidobacterium/Bacteroides ratio. Our research shows the differences in the abundances of the two enzymes that are crucial for the utilization of short type 1 oligosaccharides.IMPORTANCE Knowing the limitations of taxonomy-based research, there is an emerging need for the development of higher-resolution techniques. The significance of this research is demonstrated by the novel method used for the analysis of function-level metagenomes. BiomeScout-the presented technology-utilizes proprietary algorithms for the detection of differences between functionalities present in metagenomic samples.