We recently developed a graph deep learning method that considers contextual histopathological features from the whole-slide images. We show that the proposed method can provide interpretable prognostic biomarkers in a semi-supervised manner. We believe it will aid prognostic tasks in the future.

We recently developed oligonucleotide libraries purification technology called "MOPSS". We repurposed next generation sequencing (NGS) instrument for massively parallel purification of oligonucleotides libraries and we believe that MOPSS will flourish nucleic acid-based technologies.