Since a few years ago, we have been working in the development of computational tools for the prediction of pathological mutations in proteins (PMID: 11812146, 15390262, 15879453, 16208716, 17059831). In this work we have followed a two-step approach. First, we have characterized pathological mutations in terms of sequence, structure and evolutionary properties (PMID: 11812146). Second, we have used the results of this descriptive work to identify those parameters with the best predictive power. Also we have used these results to train a neural network and obtain an empirical model that allows the identification of disease-causing mutations with moderately high accuracies (around 70%; PMID: 15390262, 15879453, 16208716).

At present, we are applying this approach to the case of specific diseases. We believe that this is a natural step towards the advancement of personalized medicine. Within this context, we are characterizing the impact on the structure and function of alpha-galactosidase of Fabry disease-causing mutations. Again our approach is the same as in the general case, but here we are benefitting from the close collaboration with the group of Dr. Joan Montaner. 

Single-cell techniques can characterize cells in terms of different molecular properties and, usually, at a genome-wide scale, e.g., the gene-expression pattern of the entire genome, epigenetic marks, etc. A distinguishing characteristic of these techniques is their intrinsically large-scale nature, which ties their clinical application in biomedicine to the development of data analysis and AI techniques. In this project, we are interested in scRNAseq for two main reasons. First, it allows a simple approach to close the gap between molecular and cell-level understanding of the hereditary disease. Additionally, since scRNAseq was one of the first single-cell techniques to be developed, many datasets are publicly available (Moreno et al., 2022). This opens the door to addressing a biomedically relevant problem (Duan et al., 2020; Lähneman et al., 2020): the understanding of cell identity in molecular terms, using machine learning techniques. The resulting conceptual framework can be applied to relevant challenges in the clinical analysis of scRNAseq experiments, e.g., the assignment of cell identities in breast tumor samples. In this project, we focus on this problem, devising an original approach to bridge the gap between interactome and cellular identity.

Various studies show how the loss of specific interactions/network disruptions, contribute to the origin of hereditary disease (Luck et al., 2020; Sahni et al., 2015; Cheng et al., 2021). That is, it is relatively straightforward to relate the pathogenic impact of variants with changes in protein interaction patterns. In parallel, descriptions of the human interactome are becoming available (Luck et al., 2020) together with results confirming the intimate relationship between interactome and cell identity (Ye et al., 2018; Mohammady et al., 2019). We propose that this wealth of data enables the development of an innovative, data-driven model. This model would relate interactome and cell identity and could be applied to single-cell studies of breast tissue in both healthy individuals and cancer patients.

BIBLIOGRAPHY

.- Cheng, F., Zhao, J., Wang, Y., Lu, W., Liu, Z., Zhou, Y., Martin, W.R., Wang, R., Huang, J., Hao, T., et al. (2021). Comprehensive characterization of protein–protein interactions perturbed by disease mutations. Nat. Genet. 53, 342–353.

.- Duan, B., Zhu, C., Chuai, G., Tang, C., Chen, X., Chen, S., Fu, S., and Liu, Q. (2020). Learning for single-cell assignment. Sci. Adv. 6, eabd0855.

.- Lähnemann, D., Köster, J., Szczurek, E., McCarthy, D.J., Hicks, S.C., Robinson, M.D., Vallejos, C.A., Campbell, K.R., Beerenwinkel, N., Mahfouz, A., et al. (2020). Eleven grand challenges in single-cell data science (Genome Biology).

.- Luck, K., Kim, D.K., Lambourne, L., Spirohn, K., Begg, B.E., Bian, W., Brignall, R., Cafarelli, T., Campos-Laborie, F.J., Charloteaux, B., et al. (2020). A reference map of the human binary protein interactome. Nature 580, 402–408.

.-  Mohammadi, S., Davila-Velderrain, J., and Kellis, M. (2019). Reconstruction of cell-type specific interactomes at single-cell resolution. Cell Syst. 9, 559–568.

.- Moreno, P., Fexova, S., George, N., Manning, J.R., Miao, Z., Mohammed, S., MuñozPomer, A., Fullgrabe, A., Bi, Y., Bush, N., et al. (2022). Expression Atlas update: Gene and protein expression in multiple species. Nucleic Acids Res. 50, D129–D140.

.- Sahni, N., Yi, S., Taipale, M., Fuxman Bass, J.I., Coulombe-Huntington, J., Yang, F., Peng, J., Weile, J., Karras, G.I., Wang, Y., et al. (2015). Widespread macromolecular interaction perturbations in human genetic disorders. Cell 161, 647–660.

.- Ye, Z., and Sarkar, C.A. (2018). Towards a quantitative understanding of cell identity. Trends Cell Biol. 28, 1030–1048.