Raw data for MSI prediction from HE

• Network_comparison.xlsx contains raw data for the comparison of different neural network models on the tumor detection benchmark task. This is related to Extended Data Figure 1 in our manuscript "Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer"
• TCGA_GI_Test_set_predictions.xlsx contains predicted MSIness for test set patients in TCGA-CRC-KR, TCGA-CRC-DX, TCGA-STAD. This is related to Extended Data Figure 2 in our manuscript "Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer"
• DACHS_predictions.xlsx contains predicted MSIness for DACHS patients. This is related to Extended Data Figure 2 in our manuscript "Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer"
• DACHS_MSIness_in_geneticMSS.xlsx contains MSIness raw data for genetic MSS DACHS patients. This is related to Extended Data Figure 2 in our manuscript "Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer"

in all files, each line is one patient.