Motivation Quantitative structure-activity relationship (QSAR) analysis is commonly used in drug discovery. Collaborations among pharmaceutical institutions can lead to a better performance in QSAR prediction, however, intellectual property and related financial interests remain substantially hindering inter-institutional collaborations in QSAR modeling for drug discovery. Results For the first time, we verified the feasibility of applying the horizontal federated learning (HFL), which is a recently developed collaborative and privacy-preserving learning framework to perform QSAR analysis. A prototype platform of federated-learning-based QSAR modeling for collaborative drug discovery, i.e. FL-QSAR, is presented accordingly. We first compared the HFL framework with a classic privacy-preserving computation framework, i.e. secure multiparty computation to indicate its difference from various perspective. Then we compared FL-QSAR with the public collaboration in terms of QSAR modeling. Our extensive experiments demonstrated that (i) collaboration by FL-QSAR outperforms a single client using only its private data, and (ii) collaboration by FL-QSAR achieves almost the same performance as that of collaboration via cleartext learning algorithms using all shared information. Taking together, our results indicate that FL-QSAR under the HFL framework provides an efficient solution to break the barriers between pharmaceutical institutions in QSAR modeling, therefore promote the development of collaborative and privacy-preserving drug discovery with extendable ability to other privacy-related biomedical areas. Availability and implementation The source codes of FL-QSAR are available on the GitHub: https://github.com/bm2-lab/FL-QSAR. Supplementary information Supplementary data are available at Bioinformatics online.

ObjectivesFamilial aggregation of primary Sjögren’s syndrome (pSS), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA) and co-aggregation of these autoimmune diseases (ADs) (also called familial autoimmunity) is well recognised. However, the genetic predisposition variants that explain this clustering remains poorly defined.MethodsWe used whole-exome sequencing on 31 families (9 pSS, 11 SLE, 6 RA and 5 mixed autoimmunity), followed by heterozygous filtering and cosegregation analysis of a family-focused approach to document rare variants predicted to be pathogenic by in silico analysis. Potential importance in immune-related processes, gene ontology, pathway enrichment and overlap analyses were performed to prioritise gene sets.ResultsA range from 1 to 50 rare possible pathogenic variants, including 39 variants in immune-related genes across SLE, RA and pSS families, were identified. Among this gene set, regulation of T cell activation (p=4.06×10−7) and T cell receptor (TCR) signalling pathway (p=1.73×10−6) were particularly concentrated, including PTPRC (CD45), LCK, LAT–SLP76 complex genes (THEMIS, LAT, ITK, TEC, TESPA1, PLCL1), DGKD, PRKD1, PAK2 and NFAT5, shared across 14 SLE, RA and pSS families. TCR-interactive genes P2RX7, LAG3, PTPN3 and LAX1 were also detected. Overlap analysis demonstrated that the antiviral immunity gene DUS2 variant cosegregated with SLE, RA and pSS phenotypes in an extended family, that variants in the TCR-pathway genes CD45, LCK and PRKD1 occurred independently in three mixed autoimmunity families, and that variants in CD36 and VWA8 occurred in both RA-pSS and SLE-pSS families.ConclusionsOur preliminary results define common genetic characteristics linked to familial pSS, SLE and RA and highlight rare genetic variations in TCR signalling pathway genes which might provide innovative molecular targets for therapeutic interventions for those three ADs.

Efficient single-cell assignment without prior marker gene annotations is essential for single-cell sequencing data analysis. Current methods, however, have limited effectiveness for distinct single-cell assignment. They failed to achieve a well-generalized performance in different tasks because of the inherent heterogeneity of different single-cell sequencing datasets and different single-cell types. Furthermore, current methods are inefficient to identify novel cell types that are absent in the reference datasets. To this end, we present scLearn, a learning-based framework that automatically infers quantitative measurement/similarity and threshold that can be used for different single-cell assignment tasks, achieving a well-generalized assignment performance on different single-cell types. We evaluated scLearn on a comprehensive set of publicly available benchmark datasets. We proved that scLearn outperformed the comparable existing methods for single-cell assignment from various aspects, demonstrating state-of-the-art effectiveness with a reliable and generalized single-cell type identification and categorizing ability.

No abstract

Efficient single-cell assignment is essential for single-cell sequencing data analysis. With the explosive growth of single-cell sequencing data, multiple single-cell sequencing data sources are available for the same kind of tissue, which can be integrated to further improve single-cell assignment; however, an efficient integration strategy is still lacking due to the great challenges of data heterogeneity existing in multiple references. To this end, we present mtSC, a flexible single-cell assignment framework that integrates multiple references based on multitask deep metric learning designed specifically for cell type identification within tissues with multiple single-cell sequencing data as references. We evaluated mtSC on a comprehensive set of publicly available benchmark datasets and demonstrated its state-of-the-art effectiveness for integrative single-cell assignment with multiple references.

Virtual Touch tissue quantification provides a promising noninvasive strategy for differentiation of benign and malignant superficial lymph node lesions.

Objective Idiopathic inflammatory myositis-associated interstitial lung disease (IIM-ILD) significantly increases morbidity and mortality. Lung ultrasound B-lines and Krebs von den Lungen-6 (KL-6) are identified as new sonographic and serum markers of ILD, respectively. The aim of our work was to assess the role of B-lines and KL-6 as markers of the severity of IIM-ILD. For this purpose, the correlation among B-lines score, serum KL-6 levels, high-resolution CT (HRCT) score, and pulmonary function tests were investigated in IIM-ILD patients. Methods Thirty-eight patients with IIM-ILD underwent chest HRCT scans, lung ultrasound and pulmonary function tests (independently performed within 1 week) examination. To assess severity and extent of ILD at HRCT, the Warrick score was used. The B-lines score denoting the extension of ILD was calculated by summing the number of B-lines on a total of 50 scanning sites. Serum KL-6 levels (U/ml) was measured by chemiluminescent enzyme immunoassay. Results A significant correlation was found between the B-lines score and serum KL-6 levels (r = 0.43, P < 0.01), and between the Warrick score and serum KL-6 levels (r = 0.45, P < 0.01). A positive correlation between B-lines score and the Warrick score (r = 0.87, P < 0.0001) was also confirmed. Both B-lines score and KL-6 levels inversely correlated to diffusion capacity for carbon monoxide (r = −0.77, P < 0.0001 and r = −0.42, P < 0.05, respectively) and total lung capacity (r = −0.73, P < 0.0001 and r = −0.36, P < 0.05, respectively). Moreover, B-lines correlated inversely with forced vital capacity (r = −0.73, P < 0.0001), forced expiratory volume in 1 s (r = −0.69, P < 0.0001). Conclusion B-lines score and serum KL-6 levels correlate with HRCT findings and pulmonary function tests, supporting their use as measures of IIM-ILD severity.

The VTQ technique might be a useful noninvasive strategy for assessment of salivary glands in the early stage of primary Sjögren syndrome.