Search

Product-specific guidances (PSGs) recommended by the United States Food and Drug Administration (FDA) are instrumental to promote and guide generic drug product development. To assess a PSG, the FDA assessor needs to take extensive time and effort to manually retrieve supportive drug information of absorption, distribution, metabolism, and excretion (ADME) from the reference listed drug labeling. In this work, we leveraged the state-of-the-art pre-trained language models to automatically label the ADME paragraphs in the pharmacokinetics section from the FDA-approved drug labeling to facilitate PSG assessment. We applied a transfer learning approach by fine-tuning the pre-trained Bidirectional Encoder Representations from Transformers (BERT) model to develop a novel application of ADME semantic labeling, which can automatically retrieve ADME paragraphs from drug labeling instead of manual work. We demonstrate that fine-tuning the pre-trained BERT model can outperform conventional machine learning techniques, achieving up to 12.5% absolute F1 improvement. To our knowledge, we were the first to successfully apply BERT to solve the ADME semantic labeling task. We further assessed the relative contribution of pre-training and fine-tuning to the overall performance of the BERT model in the ADME semantic labeling task using a series of analysis methods, such as attention similarity and layer-based ablations. Our analysis revealed that the information learned via fine-tuning is focused on task-specific knowledge in the top layers of the BERT, whereas the benefit from the pre-trained BERT model is from the bottom layers., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 Elsevier Inc. All rights reserved.)

The psychrophilic (cold-loving) fungus Pseudogymnoascus destructans was discovered more than a decade ago to be the pathogen responsible for white-nose syndrome, an emerging disease of North American bats causing unprecedented population declines. The same species of fungus is found in Europe but without associated mortality in bats. We found P. destructans was infected with a mycovirus [named Pseudogymnoascus destructans partitivirus 1 (PdPV-1)]. The virus is bipartite, containing two double-stranded RNA (dsRNA) segments designated as dsRNA1 and dsRNA2. The cDNA sequences revealed that dsRNA1 dsRNA is 1,683 bp in length with an open reading frame (ORF) that encodes 539 amino acids (molecular mass of 62.7 kDa); dsRNA2 dsRNA is 1,524 bp in length with an ORF that encodes 434 amino acids (molecular mass of 46.9 kDa). The dsRNA1 ORF contains motifs representative of RNA-dependent RNA polymerase (RdRp), whereas the dsRNA2 ORF sequence showed homology with the putative capsid proteins (CPs) of mycoviruses. Phylogenetic analyses with PdPV-1 RdRp and CP sequences indicated that both segments constitute the genome of a novel virus in the family Partitiviridae. The purified virions were isometric with an estimated diameter of 33 nm. Reverse transcription PCR (RT-PCR) and sequencing revealed that all US isolates and a subset of Czech Republic isolates of P. destructans were infected with PdPV-1. However, PdPV-1 appears to be not widely dispersed in the fungal genus Pseudogymnoascus, as non-pathogenic fungi P. appendiculatus (1 isolate) and P. roseus (6 isolates) tested negative. P. destructans PdPV-1 could be a valuable tool to investigate fungal biogeography and the host-pathogen interactions in bat WNS.

There are several drug products that bind phosphate or bile acid in the gastrointestinal (GI) tract to exert their therapeutic efficacy. In vitro binding studies are used to assess bioequivalence (BE) of these products. The objective of this study is to identify the common deficiencies in Abbreviated New Drug Applications (ANDAs) for these products. Deficiencies were compiled from ANDAs containing in vitro binding BE studies. The deficiencies were classified into eight categories: Pre-Study Method Validation, During-Study Sample Analysis, Study Design, Study Procedure, Dissolution/Disintegration, Analytical Site Inspection, Data Submission, and Formulations. Within each category, additional subcategories were defined to characterize the deficiencies. A total of 712 deficiencies from 95 ANDAs for 11 drug products were identified and included in the analysis. The four categories with the most deficiencies were During-Study Sample Analysis (27.8%), Pre-Study Method Validation (17.3%), Data Submission (16.7%), and Study Design (15.7%). For the During-Study Sample Analysis category, failure to submit complete raw data or analytical runs ranked as the top deficiency (32.8%). For the Study Design category, using an unacceptable alternate study design (26.8%) was the most common deficiency. Within this category, other commonly occurring deficiencies included incorrect/insufficient number of absorbent concentrations, failure to pre-treat drug product with acid, insufficient number of replicates in study, incorrect calculation of k1 and k2 values, incorrect dosage form or pooled samples used in the study, and incorrect pH of study medium. The review and approval of these products may be accelerated if these common deficiencies are addressed in the original ANDA submissions.