BERMAD: batch effect removal for single-cell RNA-seq data using a multi-layer adaptation autoencoder with dual-channel framework.

MOTIVATION: Removal of batch effect between multiple datasets from different experimental platforms has become an urgent problem, since single-cell RNA sequencing (scRNA-seq) techniques developed rapidly. Although there have been some methods for this problem, most of them still face the challenge of under-correction or over-correction. Specifically, handling batch effect in highly nonlinear scRNA-seq data requires a more powerful model to address under-correction. In the meantime, some previous methods focus too much on removing difference between batches, which may disturb the biological signal heterogeneity of datasets generated from different experiments, thereby leading to over-correction. RESULTS: In this article, we propose a novel multi-layer adaptation autoencoder with dual-channel framework to address the under-correction and over-correction problems in batch effect removal, which is called BERMAD and can achieve better results of scRNA-seq data integration and joint analysis. First, we design a multi-layer adaptation architecture to model distribution difference between batches from different feature granularities. The distribution matching on various layers of autoencoder with different feature dimensions can result in more accurate batch correction outcome. Second, we propose a dual-channel framework, where the deep autoencoder processing each single dataset is independently trained. Hence, the heterogeneous information that is not shared between different batches can be retained more completely, which can alleviate over-correction. Comprehensive experiments on multiple scRNA-seq datasets demonstrate the effectiveness and superiority of our method over the state-of-the-art methods. AVAILABILITY AND IMPLEMENTATION: The code implemented in Python and the data used for experiments have been released on GitHub (https://github.com/zhanglabNKU/BERMAD) and Zenodo (https://zenodo.org/records/10695073) with detailed instructions.

Anatomical partition of the clival region and adjacent structures via extended endoscopic endonasal approach.

OBJECTIVES: To explore the anatomy of the ventral clivus and adjacent structure in the endoscopic surgery through the anterior approach, particularly in accurate locating lesions in transnasal endoscopic surgery. PATIENTS AND METHODS: A total of 9 formalin-fixed adult cadaver heads were injected with red and blue latex to observe the arteries and veins, respectively. The relationships between various parts of internal carotid artery (ICA) and anatomic structures of clivus were investigated, followed by the measurement of the posterior pharyngeal wall, anterior wall and posterior wall of clivus, cerebral dura mater, subdural space and adjacent regions to determine their correlations, as well as the clivus and adjacent structures. RESULTS: The clivus structure was divided into the bone segment, the ICA segment and subdural segment for anatomic division according to the anatomic landmarks in the anatomic process. The clivus can be classified in a shape of '' with the ICA, including the middle superior region, middle inferior region, bilateral lateral superior and lateral inferior regions. CONCLUSION: The ICA is closely related to the ventral clivus and adjacent structure, which can be used as the basis of anatomic division via anterior approach.

Endoscopic approach to the trigeminal nerve: an anatomic study.

OBJECTIVE: To describe an endoscopic perspective of the surgical anatomy of the trigeminal nerve. METHODS: Nine adult cadaveric heads were dissected endoscopically. RESULTS: Opening the pterygopalatine fossa is important because many key anatomical structures (V2, pterygopalatine ganglion, vidian nerve) can be identified and traced to other areas of the trigeminal nerve. From the pterygopalatine ganglion, the maxillary nerve and vidian nerve can be identified, and they can be traced to the gasserian ganglion and internal carotid artery. An anteromedial maxillectomy increases the angle of approach from the contralateral nares due to an increase in diameter of the piriform aperture, and provides excellent access to the mandibular nerve, the petrous carotid, and the cochlea. CONCLUSIONS: Identification of key anatomical structures in the pterygopalatine fossa can be used to identify other areas of the trigeminal nerve, and an anteromedial maxillectomy is necessary to expose the ipsilateral mandibular nerve and contralateral cranial level of the trigeminal nerve.