Source author record

Xiaoyang Fang

Xiaoyang Fang appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

cond-mat.mtrl-sci Machine Learning physics.app-ph physics.atm-clus Quantitative Methods

Catalog footprint

What is connected

3works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Subtype-Former: a deep learning approach for cancer subtype discovery with multi-omics data

Motivation: Cancer is heterogeneous, affecting the precise approach to personalized treatment. Accurate subtyping can lead to better survival rates for cancer patients. High-throughput technologies provide multiple omics data for cancer subtyping. However, precise cancer subtyping remains challenging due to the large amount and high dimensionality of omics data. Results: This study proposed Subtype-Former, a deep learning method based on MLP and Transformer Block, to extract the low-dimensional representation of the multi-omics data. K-means and Consensus Clustering are also used to achieve accurate subtyping results. We compared Subtype-Former with the other state-of-the-art subtyping methods across the TCGA 10 cancer types. We found that Subtype-Former can perform better on the benchmark datasets of more than 5000 tumors based on the survival analysis. In addition, Subtype-Former also achieved outstanding results in pan-cancer subtyping, which can help analyze the commonalities and differences across various cancer types at the molecular level. Finally, we applied Subtype-Former to the TCGA 10 types of cancers. We identified 50 essential biomarkers, which can be used to study targeted cancer drugs and promote the development of cancer treatments in the era of precision medicine.

preprint2020arXiv

A Compact Design of Four-degree-of-freedom Transmission Electron Microscope Holder for Quasi-Four-Dimensional Characterization

Electron tomography (ET) has been demonstrated to be a powerful tool in addressing challenging problems, such as understanding 3D interactions among various microstructures. Advancing ET to broader applications requires novel instrumentation design to break the bottlenecks both in theory and in practice. In this work, we built a compact four-degree-of-freedom (three-directional positionings plus self-rotation) nano-manipulator dedicated to ET applications, which is called X-Nano transmission electron microscope (TEM) holder. All the movements of the four degrees of freedom are precisely driven by built-in piezoelectric actuators, minimizing the artefacts due to the vibration and drifting of the TEM stage. Full 360o rotation is realized with an accuracy of 0.05o in the whole range, which solves the missing wedge problem. Meanwhile, the specimen can move to the rotation axis with an integrated 3D nano-manipulator, greatly reducing the effort in tracking sample locations during tilting. Meanwhile, in-situ stimulation function can be seamlessly integrated into the X-Nano TEM holder so that dynamic information can be uncovered. We expect that more delicate researches, such as those about 3D microstructural evolution, can be carried out extensively by means of this holder in the near future.

preprint2016arXiv

Dislocation Activities at the Martensite Phase Transformation Interface in Metastable Austenitic Stainless Steel: An In-situ TEM Study

Understanding the mechanism of martensitic transformation is of great importance in developing advanced high strength steels, especially TRansformation-Induced Plasticity (TRIP) steels. The TRIP effect leads to enhanced work-hardening rate, postponed onset of necking and excellent formability. In-situ transmission electron microscopy has been performed to systematically investigate the dynamic interactions between dislocations and alpha martensite at microscale. Local stress concentrations, e.g. from notches or dislocation pile-ups, render free edges and grain boundaries favorable nucleation sites for alpha martensite. Its growth leads to partial dislocation emission on two independent slip planes from the hetero-interface when the austenite matrix is initially free of dislocations. The kinematic analysis reveals that activating slip systems on two independent {111} planes of austenite are necessary in accommodating the interfacial mismatch strain. Full dislocation emission is generally observed inside of austenite regions that contain high density of dislocations. In both situations, phase boundary propagation generates large amounts of dislocations entering into the matrix, which renders the total deformation compatible and provide substantial strain hardening of the host phase. These moving dislocation sources enable plastic relaxation and prevent local damage accumulation by intense slipping on the softer side of the interfacial region. Thus, finely dispersed martensite distribution renders plastic deformation more uniform throughout the austenitic matrix, which explains the exceptional combination of strength and ductility of TRIP steels.