Geraldes, A. | Search bioRxiv

bioRxiv2022

Large-scale systematic feasibility study on the pan-cancer predictability of multi-omic biomarkers from whole slide images with deep learning

We assessed the pan-cancer predictability of multi-omic biomarkers from haematoxylin and eosin (H&E)-stained whole slide images (WSI) using deep learning (DL) throughout a systematic study. A total of 13,443 DL models predicting 4,481 multi-omic biomarkers across 32 cancer types were trained and validated. The investigated biomarkers included a broad range of genetic, transcriptomic, proteomic, and metabolic alterations, as well as established markers relevant for prognosis, molecular subtypes and clinical outcomes. Overall, we found that DL can predict multi-omic biomarkers directly from routine histology images across solid cancer types, with 50% of the models performing at an area under the curve (AUC) of more than 0.633 (with 25% of the models having an AUC larger than 0.711). A wide range of biomarkers were detectable from routine histology images across all investigated cancer types, with a mean AUC of at least 0.62 in almost all malignancies. Strikingly, we observed that biomarker predictability was mostly consistent and not dependent on sample size and class ratio, suggesting a degree of true predictability inherent in histomorphology. Together, the results of our study show the potential of DL to predict a multitude of biomarkers across the omics spectrum using only routine slides. This paves the way for accelerating diagnosis and developing more precise treatments for cancer patients.

cancer biology↗

bioRxiv2022

Evaluation of a predictive method for the H&E-based molecular profiling of breast cancer with deep learning

We present a public validation of PANProfiler (ER, PR, HER2), an in-vitro medical device (IVD) that predicts the qualitative status of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) by analysing the hematoxylin and eosin (H&E)-stained tissue scan. In public validation on 648 (ER), 648 (PR) and 560 (HER2) unseen cases with known biomarker status, the device achieves an accuracy of 87% (ER), 83% (PR) and 87% (HER2). The validation offers early evidence of the ability to predict clinically relevant breast biomarkers from an H&E slide in a relevant clinical setting.

cancer biology↗

bioRxiv2018

Population genomic analyses reveal a highly differentiated genetic cluster of northern goshawks (Accipiter gentilis laingi) in Haida Gwaii

Accurate knowledge of geographic ranges and genetic relationships among populations is important when managing a species or population of conservation concern. In the western Canadian province of British Columbia, a subspecies of the northern goshawk (Accipiter gentilis laingi) is designated as Threatened under the Canadian Species at Risk Act. Historically, the range of this bird of prey has been ambiguous and its genetic distinctness from the other North American subspecies (Accipiter gentilis atricapillus) has not been well established. Given the uncertainty in using morphological traits to assign individual goshawks to these two subspecies, we analyzed genomic relationships in tens of thousands of single nucleotide polymorphisms identified using genotyping-bysequencing of high-quality genetic samples. This genome-wide analysis revealed a genetically distinct population of northern goshawks on the archipelago of Haida Gwaii and subtle genetic structuring among the remainder of our sampling sites within North America. Following from this analysis, we developed targeted genotyping assays for ten loci that are highly differentiated between the two main genetic clusters, allowing the addition of hundreds of low-quality samples to our analysis. This additional information confirmed that the distinct genetic cluster on Haida Gwaii is restricted to that archipelago. As the laingi form was originally described as being based in Haida Gwaii, where the type specimen of that form is from, further study (especially of morphological traits) may indicate a need to restrict this name to the Haida Gwaii genetic cluster. Regardless of taxonomic treatment, our finding of a distinct Haida Gwaii genetic cluster along with the small and historically declining population size of the Haida Gwaii population suggests a high risk of extinction of an ecologically and genetically distinct form of northern goshawk. Outside of Haida Gwaii, sampling regions along the coast of BC and southeast Alaska (often considered regions inhabited by laingi) show some subtle differentiation from other North American regions. We anticipate that these results will increase the effectiveness of conservation management of northern goshawks in northwestern North America. More broadly, other conservation-related studies of genetic variation may benefit from the two-step approach we employed that first surveys genomic variation using high-quality samples and then genotypes low-quality samples at particularly informative loci.

genomics↗