Pitfalls in re-analysis of observational omics studies: a post-mortem of the human pathology atlas.
Gilis, J.; Taelman, S.; Davey, L.; Martens, L.; Clement, L.
Show abstract
Uhlen et al. (Reports, 18 august 2017) published an open-access resource with cancer-specific marker genes that are prognostic for patient survival in seventeen different types of cancer. However, their data analysis workflow is prone to the accumulation of false positives. A more reliable workflow with flexible Cox proportional hazards models employed on the same data highlights three distinct problems with such large-scale, publicly available omics datasets from observational studies today: (i) re-analysis results can not necessarily be taken forward by others, highlighting a need to cross-check important analyses with high impact outcomes; (ii) current methods are not necessarily optimal for the re-analysis of such data, indicating an urgent need to develop more suitable methods; and (iii) the limited availability of potential confounders in public metadata renders it very difficult (if not impossible) to adequately prioritize clinically relevant genes, which should prompt an in-depth discussion on how such information could be made more readily available while respecting privacy and ethics concerns.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Survival analysis of pathway activity as a prognostic determinant in breast cancer 95%
- Efficient and Flexible Integration of Variant Characteristics in Rare Variant Association Studies Using Integrated Nested Laplace Approximation 93%
- Tumor cell intrinsic and extrinsic features predicts prognosis in estrogen receptor positive breast cancer 93%
Similar papers in this journal
- Prognostic importance of splicing-triggered aberrations of protein complex interfaces in cancer 93%
- CRUX, a platform for visualising, exploring and analysing cancer genome cohort data 92%
- Two-stage biologically interpretable neural-network models for liver cancer prognosis prediction using histopathology and transcriptomic data 91%
Similar papers in this journal
- NIMBus: a Negative Binomial Regression based Integrative Method for Mutation Burden Analysis 91%
- MADloy: Robust detection of mosaic loss of chromosome Y from genotype-array-intensity data 91%
- annoFuse: an R Package to annotate, prioritize, and interactively explore putative oncogenic RNA fusions 91%
Similar papers in this journal
- Spatial Transcriptomics Inferred from Pathology Whole-Slide Images Links Tumor Heterogeneity to Survival in Breast and Lung Cancer 92%
- Accurate Prediction of Breast Cancer Survival through Coherent Voting Networks with Gene Expression Profiling 92%
- Genome-wide investigation of gene-cancer associations for the prediction of novel therapeutic targets in oncology 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.