Biology subjects

Quirke, P.

Publications and source records attributed to Quirke, P..

4 recordsLinked to original sources

bioRxiv2022

Whole genome sequencing of 2,023 colorectal cancers reveals mutational landscapes, new driver genes and immune interactions

To characterise the somatic alterations in colorectal cancer (CRC), we conducted whole-genome sequencing analysis of 2,023 tumours. We provide the most detailed high-resolution map to date of somatic mutations in CRC, and demonstrate associations with clinicopathological features, in particular location in the large bowel. We refined the mutational processes and signatures acting in colorectal tumorigenesis. In analyses across the sample set or restricted to molecular subtypes, we identified 185 CRC driver genes, of which 117 were previously unreported. New drivers acted in various molecular pathways, including Wnt (CTNND1, AXIN1, TCF3), TGF-{beta}/BMP (TGFBR1) and MAP kinase (RASGRF1, RASA1, RAF1, and several MAP2K and MAP3K loci). Non-coding drivers included intronic neo-splice site alterations in APC and SMAD4. Whilst there was evidence of an excess of mutations in functionally active regions of the non-coding genome, no specific drivers were called with high confidence. Novel recurrent copy number changes included deletions of PIK3R1 and PWRN1, as well as amplification of CCND3 and NEDD9. Putative driver structural variants included BRD4 and SOX9 regulatory elements, and ACVR2A and ANKRD11 hotspot deletions. The frequencies of many driver mutations, including somatic Wnt and Ras pathway variants, showed a gradient along the colorectum. The Pks-pathogenic E. coli signature and TP53 mutations were primarily associated with rectal cancer. A set of unreported immune escape driver genes was found, primarily in hypermutated CRCs, most of which showed evidence of genetic evasion of the anti-cancer immune response. About 25% of cancers had a potentially actionable mutation for a known therapy. Thirty-three of the new driver genes were predicted to be essential, 17 possessed a druggable structure, and nine had a bioactive compound available. Our findings provide further insight into the genetics and biology of CRC, especially tumour subtypes defined by genomic instability or clinicopathological features.

genomics↗

bioRxiv2022

Analysis of an Indian colorectal cancer faecal microbiome collection demonstrates universal colorectal cancer-associated patterns, but closest correlation with other Indian cohorts.

It is increasingly being recognised that changes in the gut microbiome have either a causative or associative relationship with colorectal cancer (CRC). However, most of this research has been carried out in a small number of developed countries with high CRC incidence. It is unknown if lower incidence countries such as India have similar microbial associations. Having previously established protocols to facilitate microbiome research in regions with developing research infrastructure, we have now collected and sequenced microbial samples from a larger cohort study of 46 Indian CRC patients and 43 healthy volunteers. When comparing to previous global collections, these samples resemble other Asian samples, with relatively high levels of Prevotella. Predicting cancer status between cohorts shows good concordance. When compared to a previous collection of Indian CRC patients, there was similar concordance, despite different sequencing technologies between cohorts. These results show that there does seem to be a global CRC microbiome, and that some inference between studies is reasonable. However, we also demonstrate that there is definite regional variation, with more similarities between location-matched comparisons. This emphasises the importance of developing protocols and advancing infrastructure to allow as many countries as possible to contribute to microbiome studies of their own populations. ImportanceColorectal cancer is increasing in many countries, thought to be partly due to the interaction between gut bacteria and changing diets. While it is known that populations in different parts of the world have very different gut microbiomes, the study of their role in colorectal cancer is almost exclusively based in the USA and Europe. We have previously shown that there is overlap between the colorectal cancer microbiome in multiple different countries, establishing robust protocols in the process. Here we expand that into a new Indian cohort. We show that while there are similarities between countries, by concentrating on one country, we can uncover important local patterns. This shows the value of sharing expertise and ensuring that work of this nature is possible wherever this disease occurs.

cancer biology↗

bioRxiv2021

Swarm learning for decentralized artificial intelligence in cancer histopathology

Artificial Intelligence (AI) can extract clinically actionable information from medical image data. In cancer histopathology, AI can be used to predict the presence of molecular alterations directly from routine histopathology slides. However, training robust AI systems requires large datasets whose collection faces practical, ethical and legal obstacles. These obstacles could be overcome with swarm learning (SL) where partners jointly train AI models, while avoiding data transfer and monopolistic data governance. Here, for the first time, we demonstrate the successful use of SL in large, multicentric datasets of gigapixel histopathology images comprising over 5000 patients. We show that AI models trained using Swarm Learning can predict BRAF mutational status and microsatellite instability (MSI) directly from hematoxylin and eosin (H&E)-stained pathology slides of colorectal cancer (CRC). We trained AI models on three patient cohorts from Northern Ireland, Germany and the United States of America and validated the prediction performance in two independent datasets from the United Kingdom using SL-based AI models. Our data show that SL enables us to train AI models which outperform most locally trained models and perform on par with models which are centrally trained on the merged datasets. In addition, we show that SL-based AI models are data efficient and maintain a robust performance even if only subsets of local datasets are used for training. In the future, SL can be used to train distributed AI models for any histopathology image analysis tasks, overcoming the need for data transfer and without requiring institutions to give up control of the final AI model.

bioinformatics↗

bioRxiv2019

Image-based consensus molecular subtype classification (imCMS) of colorectal cancer using deep learning

Image analysis is a cost-effective tool to associate complex features of tissue organisation with molecular and outcome data. Here we predict consensus molecular subtypes (CMS) of colorectal cancer (CRC) from standard H&E sections using deep learning. Domain adversarial training of a neural classification network was performed using 1,553 tissue sections with comprehensive multi- omic data from three independent datasets. Image-based consensus molecular subtyping (imCMS) accurately classified CRC whole-slide images and preoperative biopsies, spatially resolved intratumoural heterogeneity and provided accurate secondary calls with higher discriminatory power than bioinformatic prediction. In all three cohorts imCMS established sensible classification in CMS unclassified samples, reproduced expected correlations with (epi)genomic alterations and effectively stratified patients into prognostic subgroups. Leveraging artificial intelligence for the development of novel biomarkers extracted from histological slides with molecular and biological interpretability has remarkable potential for clinical translation.

cancer biology↗