Biology subjects

Doig, A. J.

Publications and source records attributed to Doig, A. J..

3 recordsLinked to original sources

bioRxiv2024

A Machine Learning Classifier to Identify and Prioritise Genes Associated with Cardiac Development

Congenital heart disease (CHD) is a major cause of infant mortality and presents life-long challenges to individuals living with these conditions. Genetic causes are known for only a minority of types of CHD. Discovering further genetic causes is limited by challenges in prioritising candidate CHD genes. We examined a wide range of features of mouse genes, including sequence characteristics, protein localisation and interaction data, developmental expression data and gene ontology annotations. Many features differ between cardiac development and non-cardiac genes, suggesting that these two gene types can be distinguished by their attributes. Therefore, we developed a supervised machine learning (ML) method to identify Mus musculus genes with a high probability of being involved in cardiac development. These genes, when mutated, are candidates for causing human CHD. Our classifier showed a cross-validation accuracy of 81% in detecting cardiac and non-cardiac genes. From our classifier we generated predictions of the cardiac development association status for all protein-coding genes in the mouse genome. We also cross-referenced our predictions with datasets of known human CHD genes, determining which are orthologues of predicted mouse cardiac genes. Our predicted cardiac genes have a high overlap with human CHD genes. Thus, our predictions could inform the prioritisation of genes when evaluating CHD patient sequence data for genetic diagnosis. Knowledge of cardiac developmental genes may speed up reaching a genetic diagnosis for patients born with CHD. Author SummaryCongenital heart disease arises during pregnancy when the heart has formed incorrectly. These malformations affect [~]1% of newborns. Yet, despite their frequency, the underlying causes are still not known in many cases. It is clear that genetic factors contribute to these defects, and increasingly DNA sequencing is used to attempt to determine if an individual has a genetic change causative of their condition. However, analysis of patient sequence data often reveals changes that are difficult to interpret due to a lack of knowledge of the function of the gene harbouring a sequence change. We aimed to facilitate the process of sequence evaluation by predicting which genes of unknown function are likely involved in heart formation. Our predictions agree with novel experimental evidence about genes needed for heart development. We found that when mutated, a high proportion of the predicted cardiac genes do indeed cause heart defects. This result suggests that our predictions may be informative for expanding our understanding of the genetic basis of congenital heart disease.

genomics↗

bioRxiv2021

Essentiality, Protein-Protein Interactions and Evolutionary Properties are Key Predictors for Identifying Cancer Genes Using Machine Learning

The identification of genes that may be linked to cancer is of great importance for the discovery of new drug targets. The rate at which cancer genes are being found experimentally is slow, however, due to the complexity of the identification and confirmation process, giving a narrow range of therapeutic targets to investigate and develop. One solution to this problem is to use predictive analysis techniques that can accurately identify cancer gene candidates in a timely fashion. Furthermore, the effort in identifying characteristics that are linked to cancer genes is crucial to further our understanding of this disease. These characteristics can be employed in recognising therapeutic drug targets. Here, we investigated whether certain genes properties can indicate the likelihood of it to be involved in the initiation or progression of cancer. We found that for cancer, the essentiality scores tend to be higher for cancer genes than for all protein coding human genes. A machine-learning model was developed and we found that essentiality related properties and properties arising from protein-protein interaction networks or evolution are particularly effective in predicting cancer-associated genes. We were also able to identify potential drug targets that have not been previously linked with cancer, but have the characteristics of cancer-related genes. Author SummaryMutations in numerous genes are known to be involved in cancer, yet there are undoubtedly many more to be discovered. We analysed a set of hundreds of cancer genes with the aim of finding out what makes them different from genes not known to be mutated in cancer. In particular, we found that genes that are essential for the survival of an organism are more likely to be involved in cancer. We used the gene properties that we examined to develop an artificial intelligence method that can accurately predict whether a gene is involved in cancer or not. Applying the method gives hundreds of non-cancer genes that resemble cancer genes. New discoveries of cancer genes are likely to be found within this set.

cancer biology↗

bioRxiv2020

Glycine Rich Segments Adopt Polyproline II Helices Which May Contribute to Biomolecular Condensate Formation

Many intrinsically disordered proteins contain Gly-rich regions which are generally assumed to be disordered. Such regions often form biomolecular condensates which play essential roles in organizing cellular processes. However, the bases of their formation and stability are still not completely understood. Considering NMR studies of the Gly-rich H. harveyi "snow flea" antifreeze protein, we recently proposed that Gly-rich sequences, such as the third "RGG" region of Fused in Sarcoma (FUS) protein, may adopt polyproline II helices whose association might stabilize condensates. Here, this hypothesis is tested with a polypeptide corresponding to the third RGG region of FUS. NMR spectroscopy and molecular dynamics simulations suggest that significant populations of polyproline II helix are present. These findings are corroborated in a model peptide Ac-RGGYGGRGGWGGRGGY-NH2, where a peak characteristic of polyproline II helix is observed using CD spectroscopy. Its intensity suggests a polyproline II population of 40%. This result is supported by data from FTIR and NMR spectroscopies. In the latter, NOE correlations are observed between the Tyr and Arg, and Arg and Trp side chain hydrogens, confirming that side chains spaced three residues apart are close in space. Taken together, the data are consistent with a polyproline II helix, which is bent to optimize interactions between guanidinium and aromatic moieties, in equilibrium with a statistical coil ensemble. In cells, the polyproline II population of these peptides could be augmented by binding profilin protein or SH3, WW or OCRE domains, association with RNA or assembly into polyproline II helical bundles. These results lend credence to the hypothesis that Gly-rich segments of disordered proteins may form polyproline II helices which help stabilize biomolecular condensates.

biophysics↗