AI 및 Data Analysis95 [scGPT] pre-training data sets PAPERhttps://www.nature.com/articles/s41592-024-02201-0GitHub ; Data Downloadhttps://github.com/bowang-lab/scGPT/blob/main/data/cellxgene/data_config.py scGPT/data/cellxgene/data_config.py at main · bowang-lab/scGPTContribute to bowang-lab/scGPT development by creating an account on GitHub.github.com 1. 공통 변수VERSION = "2023-05-08"• CellXGene Census 릴리스 버전MAJOR_TISSUE_LIST• 주요 조직 7종:["heart", "bl.. 2025. 5. 30. [cellxgene] Data Download https://cellxgene.cziscience.com/datasets Cellxgene Data PortalFind, download, and visually explore curated and standardized single cell datasets.cellxgene.cziscience.com 첫번째, Parkinson's disease관측치(Cells) 수: 2096155변수(Genes) 수: 17267데이터 정보더보기아래는 두 번째 .h5ad 파일(2,096,155개 관측치 × 17,267개 변수)에 포함된 주요 구성 요소를 정리한 내용입니다.1. 관측치 메타데이터 (adata.obs)컬럼 이름설명n_genes셀당 검출된 유전자 수n_counts셀당 총 읽기 수(counts)Brain_ba.. 2025. 5. 30. [F1 score] Macre vs. Micro vs. Weighted Macro F1각 클래스별 f1 (𝑓1ₖ)을 먼저 구한 뒤, 단순히 평균을 냅니다.${Macro-F1} = \frac1K\sum_{k=1}^K f1_k$→ 클래스 크기(샘플 수)에 상관없이 모든 클래스를 균등하게 다룹니다. Micro F1클래스별 True Positive, False Positive, False Negative 수를 모두 합산한 뒤 f1을 계산합니다.→ 실제로는 전체 샘플에 대한 f1 이고, 다중 분류에선 accuracy와 같은 값이 됩니다. Weighted F1각 클래스 f1ₖ에 클래스별 샘플 수(지원도, support) 를 곱한 뒤 합, 그리고 전체 샘플 수로 나눕니다.${Weighted-F1} = \sum_{k=1}^K \frac{n_k}{N} \, f1_k$→ 클래스 비율을 .. 2025. 5. 28. [DosaCNV] Deep multiple-instance learning accurately predicts gene haploinsufficiency and deletion pathogenicity PAPERhttps://pmc.ncbi.nlm.nih.gov/articles/PMC10491176/ Deep multiple-instance learning accurately predicts gene haploinsufficiency and deletion pathogenicity - PMCDetails of DosaCNV. We denote Xi as the input for the i-th deletion, which is a matrix of dimensions M × L, where M represents the number of genes and L corresponds to the number of gene-level features. Given that the number of genes .. 2025. 5. 23. Isoform Function Prediction Using a Deep Neural Network PAPERhttps://arxiv.org/abs/2208.03325 Isoform Function Prediction Using a Deep Neural NetworkIsoforms are mRNAs produced from the same gene site in the phenomenon called Alternative Splicing. Studies have shown that more than 95% of human multi-exon genes have undergone alternative splicing. Although there are few changes in mRNA sequence, They maarxiv.org PAPER REVIEWIsoform Function Prediction.. 2025. 5. 23. [DosaCNV] Deep multiple-instance learning accurately predicts gene haploinsufficiency and deletion pathogenicity PAPERhttps://pmc.ncbi.nlm.nih.gov/articles/PMC10491176/ Deep multiple-instance learning accurately predicts gene haploinsufficiency and deletion pathogenicity - PMCDetails of DosaCNV. We denote Xi as the input for the i-th deletion, which is a matrix of dimensions M × L, where M represents the number of genes and L corresponds to the number of gene-level features. Given that the number of genes .. 2025. 5. 22. 이전 1 2 3 4 5 6 ··· 16 다음
