Deep learning in preclinical antibody drug discovery and development.

Antibody drugs have become a key part of biotherapeutics. Patients suffering from various diseases have benefited from antibody therapies. However, its development process is rather long, expensive and risky. To speed up the process, reduce cost and improve success rate, artificial intelligence, especially deep learning methods, have been widely used in all aspects of preclinical antibody drug development, from library generation to hit identification, developability screening, lead selection and optimization. In this review, we systematically summarize antibody encodings, deep learning architectures and models used in preclinical antibody drug discovery and development. We also critically discuss challenges and opportunities, problems and possible solutions, current applications and future directions of deep learning in antibody drug development.

Serum albumin and cardiovascular disease: a Mendelian randomization study.

BACKGROUND: An increasing body of evidence suggests that serum albumin levels play a role in cardiovascular diseases. However, the specific causal relationship between serum albumin levels and cardiovascular disease remains partially unknown. METHODS: Mendelian randomization (MR) was employed in this study to examine potential causal relationships between instrumental variables and cardiovascular diseases. Specifically, we utilized genetic variants of serum albumin levels within the reference range as our instrumental variables. To acquire data on genetic associations with cardiovascular diseases, we sourced information from renowned genome-wide association studies such as UK BioBank, EMBL-EBI, and FinnGen. Notably, our study leveraged summary statistics from large cohorts that have been previously described. RESULTS: We explored the association between serum albumin levels and various conditions, including heart failure (HF), venous thromboembolism (VTE), stroke, atrial fibrillation (AF), coronary artery disease (CAD), type 2 diabetes (T2DM), and pulmonary heart disease (PHD). Genetically predicted serum albumin levels were associated with PHD (odds ratio = 0.737, 95% CI = 0.622 - 0.874, P < 0.001), AF (odds ratio = 0.922, 95% CI = 0.870 - 0.977, P = 0.006), VTE (odds ratio = 0.993, 95% CI = 0.991 - 0.995, P < 0.001), and Stroke (odds ratio = 0.997, 95% CI = 0.995 - 0.999, P = 0.002). However, genetically predicted serum albumin level traits were not associated with HF, CAD and T2DM. CONCLUSION: Our study demonstrates a significant association between serum albumin levels and cardiovascular disease, underscoring the crucial role of low serum albumin as a predictive factor in patients with cardiovascular disease.

ACP-Dnnel: anti-coronavirus peptides' prediction based on deep neural network ensemble learning.

The ongoing COVID-19 pandemic has caused dramatic loss of human life. There is an urgent need for safe and efficient anti-coronavirus infection drugs. Anti-coronavirus peptides (ACovPs) can inhibit coronavirus infection. With high-efficiency, low-toxicity, and broad-spectrum inhibitory effects on coronaviruses, they are promising candidates to be developed into a new type of anti-coronavirus drug. Experiment is the traditional way of ACovPs' identification, which is less efficient and more expensive. With the accumulation of experimental data on ACovPs, computational prediction provides a cheaper and faster way to find anti-coronavirus peptides' candidates. In this study, we ensemble several state-of-the-art machine learning methodologies to build nine classification models for the prediction of ACovPs. These models were pre-trained using deep neural networks, and the performance of our ensemble model, ACP-Dnnel, was evaluated across three datasets and independent dataset. We followed Chou's 5-step rules. (1) we constructed the benchmark datasets data1, data2, and data3 for training and testing, and introduced the independent validation dataset ACVP-M; (2) we analyzed the peptides sequence composition feature of the benchmark dataset; (3) we constructed the ACP-Dnnel model with deep convolutional neural network (DCNN) merged the bi-directional long short-term memory (BiLSTM) as the base model for pre-training to extract the features embedded in the benchmark dataset, and then, nine classification algorithms were introduced to ensemble together for classification prediction and voting together; (4) tenfold cross-validation was introduced during the training process, and the final model performance was evaluated; (5) finally, we constructed a user-friendly web server accessible to the public at http://150.158.148.228:5000/ . The highest accuracy (ACC) of ACP-Dnnel reaches 97%, and the Matthew's correlation coefficient (MCC) value exceeds 0.9. On three different datasets, its average accuracy is 96.0%. After the latest independent dataset validation, ACP-Dnnel improved at MCC, SP, and ACC values 6.2%, 7.5% and 6.3% greater, respectively. It is suggested that ACP-Dnnel can be helpful for the laboratory identification of ACovPs, speeding up the anti-coronavirus peptide drug discovery and development. We constructed the web server of anti-coronavirus peptides' prediction and it is available at http://150.158.148.228:5000/ .

The complete chloroplast genome sequence of Rubus hirsutus Thunb. and a comparative analysis within Rubus species.

Rubus hirsutus is a type of tonifying kidney-essence herb that belongs to the Rosaceae family, and has been commonly used to treat multiple diseases, such as polyuria, impotence, and infertility. In this study, we determined the complete chloroplast sequence of R. hirsutus and conduced a comparative analysis within the genus Rubus. The assembled chloroplast (cp.) genome is 156,380 bp in length with a GC content of 37.0% and shares a conserved quadripartite structure within the other cp. genomes in this genus. A total of 132 unique genes were annotated in the cp. genome of R. hirsutus, which contained 87 protein-coding genes, 37 tRNAs, and eight rRNAs. Seventeen duplicated genes were identified in the inverted repeats region. Furthermore, 70 simple sequence repeats and 35 long repeats were detected in total in the R. hirsutus chloroplast genome. Eight mutational hotspots were identified in the cp. genome of this species with higher nucleotide variations in non-coding regions than those of coding regions. Furthermore, the gene order, codon usage, and repeat sequence distribution were highly consistent in Rubus according to the results of a comparative analysis. A phylogenetic analysis indicated that there was a sister relationship between R. hirsutus and R. chingii. Overall, the complete chloroplast genome of R. hirsutus and the comparative analysis will help to further the evolutionary study, conservation, phylogenetic reconstruction, and development of molecular barcodes for the genus Rubus.

Molecular Insights into Small-Molecule Drug Discovery for SARS-CoV-2.

The mainstream approach to antiviral drugs against COVID-19 is to focus on key stages of the SARS-CoV-2 life cycle. The vast majority of candidates under investigation are repurposed from agents of other indications. Understanding protein-inhibitor interactions at the molecular scale will provide crucial insights for drug discovery to stop this pandemic. In this article, we summarize and analyze the most recent structural data on several viral targets in the presence of promising inhibitors for COVID-19 in the context of the perspective of modes of action (MOA) to unravel insightful mechanistic features with atomistic resolution. The targets include spike glycoprotein and various host proteases mediating the entry of the virus into the cells, viral chymotrypsin- and papain-like proteases, and RNA-dependent RNA polymerase. The main purpose of this review is to present detailed MOA analysis to inspire fresh ideas for both de novo drug design and optimization of known scaffolds to combat COVID-19.

DOTAD: A Database of Therapeutic Antibody Developability.

The development of therapeutic antibodies is an important aspect of new drug discovery pipelines. The assessment of an antibody's developability-its suitability for large-scale production and therapeutic use-is a particularly important step in this process. Given that experimental assays to assess antibody developability in large scale are expensive and time-consuming, computational methods have been a more efficient alternative. However, the antibody research community faces significant challenges due to the scarcity of readily accessible data on antibody developability, which is essential for training and validating computational models. To address this gap, DOTAD (Database Of Therapeutic Antibody Developability) has been built as the first database dedicated exclusively to the curation of therapeutic antibody developability information. DOTAD aggregates all available therapeutic antibody sequence data along with various developability metrics from the scientific literature, offering researchers a robust platform for data storage, retrieval, exploration, and downloading. In addition to serving as a comprehensive repository, DOTAD enhances its utility by integrating a web-based interface that features state-of-the-art tools for the assessment of antibody developability. This ensures that users not only have access to critical data but also have the convenience of analyzing and interpreting this information. The DOTAD database represents a valuable resource for the scientific community, facilitating the advancement of therapeutic antibody research. It is freely accessible at http://i.uestc.edu.cn/DOTAD/ , providing an open data platform that supports the continuous growth and evolution of computational methods in the field of antibody development.

AB-Amy: machine learning aided amyloidogenic risk prediction of therapeutic antibody light chains.

Over 120 FDA-approved antibody-based therapeutics are used to treat a variety of diseases.However, many candidates could fail because of unfavorable physicochemical properties. Light-chain amyloidosis is one form of aggregation that can lead to severe safety risks in clinical development. Therefore, screening candidates with a less amyloidosis risk at the early stage can not only save the time and cost of antibody development but also improve the safety of antibody drugs. In this study, based on the dipeptide composition of 742 amyloidogenic and 712 non-amyloidogenic antibody light chains, a support vector machine-based model, AB-Amy, was trained to predict the light-chain amyloidogenic risk. The AUC of AB-Amy reaches 0.9651. The excellent performance of AB-Amy indicates that it can be a useful tool for the in silico evaluation of the light-chain amyloidogenic risk to ensure the safety of antibody therapeutics under clinical development. A web server is freely available at http://i.uestc.edu.cn/AB-Amy/.

A Qualitative Signature to Identify TERT Promoter Mutant High-Risk Tumors in Low-Grade Gliomas.

Background: Telomerase reverse transcriptase promoter (TERT-p) mutation has been frequently found, but associated with contrary prognosis, in both low-grade gliomas and glioblastomas. For the low-grade gliomas (Grades II-III), TERT-p mutant patients have a better prognosis than the wildtype patients, whereas for the GBMs (Grade IV), TERT-p mutation is related to a poor prognosis. We hypothesize that there exist high-risk patients in LGGs who share GBM-like molecular features, including TERT-p mutation, and need more intensive treatment than other LGGs. A molecular signature is needed to identify these high-risk patients for an accurate and timely treatment. Methods: Using the within-sample relative expression orderings of gene pairs, we identified the gene pairs with significantly stable REOs, respectively, in both the TERT-p mutant LGGs and GBMs but with opposite directions in the two groups. These reversely stable gene pairs were used as the molecular signature to stratify the LGGs into high-risk and low-risk groups. Results: A signature consisting of 21 gene pairs was developed, which can classify LGGs into two groups with significantly different overall survival. The high-risk group has a similar genetic mutation profile and a similar survival profile as GBMs, and these high-risk tumors may progress to a more malignant state. Conclusion: The 21 gene-pair signature based on REOs is capable of identifying high-risk patients in LGGs and guiding the clinical choice for appropriate and timely intervention.

Bromine Incorporation and Suppressed Cation Rotation in Mixed-Halide Perovskites.

Engineering the composition of perovskite active layers has been critical in increasing the efficiency of perovskite solar cells (PSCs) to more than 25% in the latest reports. Partial substitutions of the monovalent cation and the halogen have been adopted in the highest-performing devices, but the precise role of bromine incorporation remains incompletely explained. Here we use quasi-elastic neutron scattering (QENS) to study, as a function of the degree of bromine incorporation, the dynamics of organic cations in triple-cation lead mixed-halide perovskites. We find that the inclusion of bromine suppresses low-energy rotations of formamidinium (FA), and we find that inhibiting FA rotation correlates with a longer-lived carrier lifetime. When the fraction of bromine approaches 0.15 on the halogen site-a composition used extensively in the PSC literature-the fraction of actively rotating FA molecules is minimized: indeed, the fraction of rotating FA is suppressed by more than 25% compared to the bromine-free perovskite.

Regulating strain in perovskite thin films through charge-transport layers.

Thermally-induced tensile strain that remains in perovskite films following annealing results in increased ion migration and is a known factor in the instability of these materials. Previously-reported strain regulation methods for perovskite solar cells (PSCs) have utilized substrates with high thermal expansion coefficients that limits the processing temperature of perovskites and compromises power conversion efficiency. Here we compensate residual tensile strain by introducing an external compressive strain from the hole-transport layer. By using a hole-transport layer with high thermal expansion coefficient, we compensate the tensile strain in PSCs by elevating the processing temperature of hole-transport layer. We find that compressive strain increases the activation energy for ion migration, improving the stability of perovskite films. We achieve an efficiency of 16.4% for compressively-strained PSCs; and these retain 96% of their initial efficiencies after heating at 85 °C for 1000 hours-the most stable wide-bandgap perovskites (above 1.75 eV) reported so far.

Enhanced optical path and electron diffusion length enable high-efficiency perovskite tandems.

Tandem solar cells involving metal-halide perovskite subcells offer routes to power conversion efficiencies (PCEs) that exceed the single-junction limit; however, reported PCE values for tandems have so far lain below their potential due to inefficient photon harvesting. Here we increase the optical path length in perovskite films by preserving smooth morphology while increasing thickness using a method we term boosted solvent extraction. Carrier collection in these films - as made - is limited by an insufficient electron diffusion length; however, we further find that adding a Lewis base reduces the trap density and enhances the electron-diffusion length to 2.3 µm, enabling a 19% PCE for 1.63 eV semi-transparent perovskite cells having an average near-infrared transmittance of 85%. The perovskite top cell combined with solution-processed colloidal quantum dot:organic hybrid bottom cell leads to a PCE of 24%; while coupling the perovskite cell with a silicon bottom cell yields a PCE of 28.2%.

Combining Efficiency and Stability in Mixed Tin-Lead Perovskite Solar Cells by Capping Grains with an Ultrathin 2D Layer.

The development of narrow-bandgap (Eg ≈ 1.2 eV) mixed tin-lead (Sn-Pb) halide perovskites enables all-perovskite tandem solar cells. Whereas pure-lead halide perovskite solar cells (PSCs) have advanced simultaneously in efficiency and stability, achieving this crucial combination remains a challenge in Sn-Pb PSCs. Here, Sn-Pb perovskite grains are anchored with ultrathin layered perovskites to overcome the efficiency-stability tradeoff. Defect passivation is achieved both on the perovskite film surface and at grain boundaries, an approach implemented by directly introducing phenethylammonium ligands in the antisolvent. This improves device operational stability and also avoids the excess formation of layered perovskites that would otherwise hinder charge transport. Sn-Pb PSCs with fill factors of 79% and a certified power conversion efficiency (PCE) of 18.95% are reported-among the highest for Sn-Pb PSCs. Using this approach, a 200-fold enhancement in device operating lifetime is achieved relative to the nonpassivated Sn-Pb PSCs under full AM1.5G illumination, and a 200 h diurnal operating time without efficiency drop is achieved under filtered AM1.5G illumination.

Efficient tandem solar cells with solution-processed perovskite on textured crystalline silicon.

Stacking solar cells with decreasing band gaps to form tandems presents the possibility of overcoming the single-junction Shockley-Queisser limit in photovoltaics. The rapid development of solution-processed perovskites has brought perovskite single-junction efficiencies >20%. However, this process has yet to enable monolithic integration with industry-relevant textured crystalline silicon solar cells. We report tandems that combine solution-processed micrometer-thick perovskite top cells with fully textured silicon heterojunction bottom cells. To overcome the charge-collection challenges in micrometer-thick perovskites, we enhanced threefold the depletion width at the bases of silicon pyramids. Moreover, by anchoring a self-limiting passivant (1-butanethiol) on the perovskite surfaces, we enhanced the diffusion length and further suppressed phase segregation. These combined enhancements enabled an independently certified power conversion efficiency of 25.7% for perovskite-silicon tandem solar cells. These devices exhibited negligible performance loss after a 400-hour thermal stability test at 85°C and also after 400 hours under maximum power point tracking at 40°C.

Contactless measurements of photocarrier transport properties in perovskite single crystals.

The remarkable properties of metal halide perovskites arising from their impressive charge carrier diffusion lengths have led to rapid advances in solution-processed optoelectronics. Unfortunately, diffusion lengths reported in perovskite single crystals have ranged widely - from 3 µm to 3 mm - for ostensibly similar materials. Here we report a contactless method to measure the carrier mobility and further extract the diffusion length: our approach avoids both the effects of contact resistance and those of high electric field. We vary the density of quenchers - epitaxially included within perovskite single crystals - and report the dependence of excited state lifetime in the perovskite on inter-quencher spacing. Our results are repeatable and self-consistent (i.e. they agree on diffusion length for many different quencher concentrations) to within ± 6%. Using this method, we obtain a diffusion length in metal-halide perovskites of 2.6 µm ± 0.1 µm.

In Situ Back-Contact Passivation Improves Photovoltage and Fill Factor in Perovskite Solar Cells.

Organic-inorganic hybrid perovskite solar cells (PSCs) have seen a rapid rise in power conversion efficiencies in recent years; however, they still suffer from interfacial recombination and charge extraction losses at interfaces between the perovskite absorber and the charge-transport layers. Here, in situ back-contact passivation (BCP) that reduces interfacial and extraction losses between the perovskite absorber and the hole transport layer (HTL) is reported. A thin layer of nondoped semiconducting polymer at the perovskite/HTL interface is introduced and it is shown that the use of the semiconductor polymer permits-in contrast with previously studied insulator-based passivants-the use of a relatively thick passivating layer. It is shown that a flat-band alignment between the perovskite and polymer passivation layers achieves a high photovoltage and fill factor: the resultant BCP enables a photovoltage of 1.15 V and a fill factor of 83% in 1.53 eV bandgap PSCs, leading to an efficiency of 21.6% in planar solar cells.

The Dawn of Lead-Free Perovskite Solar Cell: Highly Stable Double Perovskite Cs2AgBiBr6 Film.

Recently, lead-free double perovskites have emerged as a promising environmentally friendly photovoltaic material for their intrinsic thermodynamic stability, appropriate bandgaps, small carrier effective masses, and low exciton binding energies. However, currently no solar cell based on these double perovskites has been reported, due to the challenge in film processing. Herein, a first lead-free double perovskite planar heterojunction solar cell with a high quality Cs2AgBiBr6 film, fabricated by low-pressure assisted solution processing under ambient conditions, is reported. The device presents a best power conversion efficiency of 1.44%. The preliminary efficiency and the high stability under ambient condition without encapsulation, together with the high film quality with simple processing, demonstrate promise for lead-free perovskite solar cells.

Nano-structured electron transporting materials for perovskite solar cells.

Organic-inorganic hybrid perovskite solar cells have been developing rapidly in the past several years, and their power conversion efficiency has reached over 20%, nearing that of polycrystalline silicon solar cells. Because the diffusion length of the hole in perovskites is longer than that of the electron, the performance of the device can be improved by using an electron transporting layer, e.g., TiO2, ZnO and TiO2/Al2O3. Nano-structured electron transporting materials facilitate not only electron collection but also morphology control of the perovskites. The properties, morphology and preparation methods of perovskites are reviewed in the present article. A comprehensive understanding of the relationship between the structure and property will benefit the precise control of the electron transporting process and thus further improve the performance of perovskite solar cells.