A generalizable and robust deep learning algorithm for mitosis detection in multicenter breast histopathological images.

Mitosis counting of biopsies is an important biomarker for breast cancer patients, which supports disease prognostication and treatment planning. Developing a robust mitotic cell detection model is highly challenging due to its complex growth pattern and high similarities with non-mitotic cells. Most mitosis detection algorithms have poor generalizability across image domains and lack reproducibility and validation in multicenter settings. To overcome these issues, we propose a generalizable and robust mitosis detection algorithm (called FMDet), which is independently tested on multicenter breast histopathological images. To capture more refined morphological features of cells, we convert the object detection task as a semantic segmentation problem. The pixel-level annotations for mitotic nuclei are obtained by taking the intersection of the masks generated from a well-trained nuclear segmentation model and the bounding boxes provided by the MIDOG 2021 challenge. In our segmentation framework, a robust feature extractor is developed to capture the appearance variations of mitotic cells, which is constructed by integrating a channel-wise multi-scale attention mechanism into a fully convolutional network structure. Benefiting from the fact that the changes in the low-level spectrum do not affect the high-level semantic perception, we employ a Fourier-based data augmentation method to reduce domain discrepancies by exchanging the low-frequency spectrum between two domains. Our FMDet algorithm has been tested in the MIDOG 2021 challenge and ranked first place. Further, our algorithm is also externally validated on four independent datasets for mitosis detection, which exhibits state-of-the-art performance in comparison with previously published results. These results demonstrate that our algorithm has the potential to be deployed as an assistant decision support tool in clinical practice. Our code has been released at https://github.com/Xiyue-Wang/1st-in-MICCAI-MIDOG-2021-challenge.

Undescribed benzophenone and xanthones from cave-derived Streptomyces sp. CB09001.

A new benzophenone huanglongmycin (HLM) D (1) and two new monomeric xanthones huanglongmycin E (2) and F (3), together with four known aromatic polyketides aloesaponarin II (4) and the previously isolated huanglongmycin A-C (5-7) were obtained from cave-derived Streptomyces sp. CB09001. The structures of 1-3 were established based on 1D, 2D NMR and HRMS data. Compounds 1-7 may be biosynthesized by a type II huanglongmycin polyketide synthase based on gene inactivation of hlmG encoding KSɑ in hlm gene cluster and their plausible biosynthetic mechanism was proposed.

Syn-2, 3-diols and anti-inflammatory indole derivatives from Streptomyces sp. CB09001.

Two new natural diols, (2S, 3S, 4S)-4-methyl-1-phenylhexane-2,3-diol (1) and (2S, 3S)-4-methyl-1-phenylpentane-2,3-diol (2), together with five known compounds, xenocyloins B-D (3-5), lumichrome (6) and thymidine (7) were isolated from Streptomyces sp. CB09001. The absolute configurations of 1 and 2 were established by crystallographic structure analysis. The anti-inflammatory effects of 1-7 were also investigated in RAW246.7 murine macrophage cells stimulated by lipopolysaccharide. The indole derivative xenocyloin B (3) significantly inhibited inducible nitric oxide synthase expression in RAW264.7 cells and could be a potential anti-inflammatory drug lead.

A 3-hydroxy-3-methylglutaryl-CoA synthase-based probe for the discovery of the acyltransferase-less type I polyketide synthases.

Acyltransferase (AT)-less type I polyketide synthases (PKSs) produce complex natural products due to the presence of many unique tailoring enzymes. The 3-hydroxy-3-methylglutaryl coenzyme A synthases (HCSs) are responsible for ß-alkylation of the growing polyketide intermediates in AT-less type I PKSs. In this study, we discovered a large group of HCSs, closely associated with the characterized and orphan AT-less type I PKSs through in silico genome mining, sequence and genome neighbourhood network analyses. Using HCS-based probes, the survey of 1207 in-house strains and 18 soil samples from different geographic locations revealed the vast diversity of HCS-containing AT-less type I PKSs. The presence of HCSs in many AT-less type I PKSs suggests their co-evolutionary relationship. This study provides a new probe to study the abundance and diversity of AT-less type I PKSs in the environment and microbial strain collections. Our study should inspire future efforts to discover new polyketide natural products from AT-less type I PKSs.

Huanglongmycin A-C, Cytotoxic Polyketides Biosynthesized by a Putative Type II Polyketide Synthase From Streptomyces sp. CB09001.

Three natural products of nonaketide biosynthetic origin, probably biosynthesized from nine molecules of malonyl-CoA, have been isolated. Herein we described the isolation and structure elucidation of huanglongmycin (HLM) A-C and identification of the putative hlm biosynthetic gene cluster from Streptomyces sp. CB09001, isolated from a karstic cave in Xiangxi, China. Albeit previously isolated, HLM A was reported for the first time to exhibit moderate cytotoxicity against A549 lung cancer cell line (IC50 = 13.8 ± 1.5 µM) and weak antibacterial activity against gram-negative clinical isolates. A putative biosynthetic pathway for HLM A, featuring a nonaketide-specific type II polyketide synthase, was proposed. It would be consistent with the isolation of HLM B and C, which are two new natural products and likely shunt metabolites during HLM A biosynthesis.