Protein embeddings learned from aligned sequences have been leveraged in a wide array of tasks in protein understanding and engineering. The sequence embeddings are generated through semi-supervised training on millions of sequences with deep neural models defined with hundreds of millions of parameters, and they continue to increase in performance on target tasks with increasing complexity. We report a more data-efficient approach to encode protein information through joint training on protein sequence and structure in a semi-supervised manner. We show that the method is able to encode both types of information to form a rich embedding space which can be used for downstream prediction tasks. We show that the incorporation of rich structural information into the context under consideration boosts the performance of the model by predicting the effects of single-mutations. We attribute increases in accuracy to the value of leveraging proximity within the enriched representation to identify sequentially and spatially close residues that would be affected by the mutation, using experimentally validated or predicted structures.

No abstract

Pulmonary tumor embolism (PTE) and pulmonary tumor thrombotic microangiopathy (PTTM) are rare etiologies for rapidly progressive dyspnea in the setting of undiagnosed metastatic cancer. They occur most frequently in association with adenocarcinomas, with PTE being most frequently associated with hepatocellular carcinoma and PTTM being most commonly reported with gastric adenocarcinoma. Pulmonary tumor embolism and PTTM appear to be a disease spectrum where PTTM represents an advanced form of PTE. Pulmonary tumor embolism and PTTM are mostly identified postmortem during autopsy as the antemortem diagnosis remains a clinical challenge due to the rapidly progressive nature of these rare diseases. We report 2 cases of rapidly progressive respiratory failure leading to death, due to tumoral pulmonary hypertension resulting from PTE and PTTM, diagnosed postmortem. Both of the patients were middle-aged females, nonsmokers, and had a gastrointestinal source of their primary malignancy.

This study describes a subset of patients with posterior urethral valves (PUV) who presented late in childhood. The objective was to identify factors that lead to back-pressure effects on the upper tracts, which persist in spite of adequate valve ablation in some patients, and seek factors that may preserve the upper tracts despite untreated obstruction in other patients. Six children with PUV diagnosed after infancy were evaluated. The pre-operative work-up included renal biochemistry, ultrasonography, voiding cystourethrography, and uroflowmetry. Detailed urodynamic studies, including uroflowmetry and slow-fill cystometry, were performed in all cases 6 months after surgery. Adequacy of valve fulguration was confirmed by urethroscopy. Three of the six patients had normal upper tracts; in these, there was marked improvement in peak urine flow rates after fulguration and bladder pressures were normal. The other three patients had bilateral hydroureteronephrosis, and two had chronic renal failure. This group had markedly decreased functional bladder capacity with loss of compliance at low bladder volumes and significant residual urine volumes in spite of adequate valve fulguration, suggesting myogenic detrusor failure. We conclude that in patients with PUV presenting beyond the age of 5 years, upper-tract deterioration may accompany high storage pressures in the bladder. In some boys with long-standing obstruction the upper tracts may escape damage; in our series this was associated with normal bladder dynamics and appeared unrelated to the severity or duration of outflow obstruction.