Camila de Souza

Dr. Camila de Souza

orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 February 2020 in bioRxiv
Verified
Preprint
Open access
PDF file
GRGurjit RANDHAWA
Gurjit RANDHAWA
MSMaximillian Soltysiak
Maximillian Soltysiak
HRHadi El Roz
Hadi El Roz
CSCamila de Souza
Camila de Souza
KHKathleen A. Hill
Kathleen A. Hill
LKLila Kari
Lila Kari
As of February 20, 2020, the 2019 novel coronavirus (renamed to COVID-19) spread to 30 countries with 2130 deaths and more than 75500 confirmed cases. COVID-19 is being compared to the infamous SARS coronavirus, which resulted, between November 2002 and July 2003, in 8098 confirmed cases worldwide with a 9.6% death rate and 774 deaths. Though COVID-19 has a death rate of 2.8% as of 20 February, the 75752 confirmed cases in a few weeks (December 8, 2019 to February 20, 2020) are alarming, with cases likely being under-reported given the comparatively longer incubation period. Such outbreaks demand elucidation of taxonomic classification and origin of the virus genomic sequence, for strategic planning, containment, and treatment. This paper identifies an intrinsic COVID-19 genomic signature and uses it together with a machine learning-based alignment-free approach for an ultra-fast, scalable, and highly accurate classification of whole COVID-19 genomes. The proposed method combines supervised machine learning with digital signal processing for genome analyses, augmented by a decision tree approach to the machine learning component, and a Spearman’s rank correlation coefficient analysis for result validation. These tools are used to analyze a large dataset of over 5000 unique viral genomic sequences, totalling 61.8 million bp. Our results support a hypothesis of a bat origin and classify COVID-19 asSarbecovirus, withinBetacoronavirus. Our method achieves high levels of classification accuracy and discovers the most relevant relationships among over 5,000 viral genomes within a few minutes,ab initio, using raw DNA sequence data alone, and without any specialized biological knowledge, training, gene or genome annotations. This suggests that, for novel viral and pathogen genome sequences, this alignment-free whole-genome machine-learning approach can provide a reliable real-time option for taxonomic classification.

23 September 2020 in PLOS Computational Biology
Verified
Journal Article
Open access
PDF file
CSCamila de Souza
Camila de Souza
MAMirela Andronescu
Mirela Andronescu
TMTehmina Masud
Tehmina Masud
FKFarhia Kabeer
Farhia Kabeer
JBJustina Biele
Justina Biele
SSSohrab P. Shah
Sohrab P. Shah
We present Epiclomal, a probabilistic clustering method arising from a hierarchical mixture model to simultaneously cluster sparse single-cell DNA methylation data and impute missing values. Using synthetic and published single-cell CpG datasets, we show that Epiclomal outperforms non-probabilistic methods and can handle the inherent missing data characteristic that dominates single-cell CpG genome sequences. Using newly generated single-cell 5mCpG sequencing data, we show that Epiclomal discovers sub-clonal methylation patterns in aneuploid tumour genomes, thus defining epiclones that can match or transcend copy number-determined clonal lineages and opening up an important form of clonal analysis in cancer. Epiclomal is written in R and Python and is available at https://github.com/shahcompbio/Epiclomal. DNA methylation is an epigenetic mark that occurs when methyl groups are attached to the DNA molecule, thereby playing decisive roles in numerous biological processes. Advances in technology have allowed the generation of high-throughput DNA methylation sequencing data from single cells. One of the goals is to group cells according to their DNA methylation profiles; however, a major challenge arises due to a large amount of missing data per cell. To address this problem, we developed a novel statistical model and framework: Epiclomal. Our approach uses a hierarchical mixture model to borrow statistical strength across cells and neighboring loci to accurately define cell groups (clusters). We compare our approach to different methods on both synthetic and published datasets. We show that Epiclomal is more robust than other approaches, producing more accurate clusters of cells in the majority of experimental scenarios. We also apply Epiclomal to newly generated single-cell DNA methylation data from breast cancer xenografts. Our results show that methylation-based clusters can mirror or in some instances transcend the clusters defined by single-cell copy number analysis. This illustrates the importance of single-cell DNA methylation analysis in understanding cellular heterogeneity in cancer.

3 February 2023 in Journal of Applied Statistics
Verified
Journal Article
Closed access
PSPedro Henrique T. O. Sousa
Pedro Henrique T. O. Sousa
CSCamila de Souza
Camila de Souza
RDRonaldo Dias
Ronaldo Dias
Considering the context of functional data analysis, we developed and applied a new Bayesian approach via the Gibbs sampler to select basis functions for a finite representation of functional data. The proposed methodology uses Bernoulli latent variables to assign zero to some of the basis function coefficients with a positive probability. This procedure allows for an adaptive basis selection since it can determine the number of bases and which ones should be selected to represent functional data. Moreover, the proposed procedure measures the uncertainty of the selection process and can be applied to multiple curves simultaneously. The methodology developed can deal with observed curves that may differ due to experimental error and random individual differences between subjects, which one can observe in a real dataset application involving daily numbers of COVID-19 cases in Brazil. Simulation studies show the main properties of the proposed method, such as its accuracy in estimating the coefficients and the strength of the procedure to find the true set of basis functions. Despite having been developed in the context of functional data analysis, we also compared the proposed model via simulation with the well-established LASSO and Bayesian LASSO, which are methods developed for non-functional data.

5 May 2023 in International Journal of Climatology
Verified
Journal Article
Closed access
YNYeu Deck Ngui
Yeu Deck Ngui
Mohammad Reza NajafiMohammad Reza Najafi
Mohammad Reza Najafi
CSCamila de Souza
Camila de Souza
DSDavid M. L. Sills
David M. L. Sills

30 April 2024 in Advances in Data Analysis and Classification
Verified
Journal Article
Closed access
PDF file
CXChengqian Xian
Chengqian Xian
CSCamila de Souza
Camila de Souza
JJJohn Jewell
John Jewell
RDRonaldo Dias
Ronaldo Dias

Verified
Journal Article
Open access
PDF file
ZFZahra Fazel
Zahra Fazel
CSCamila de Souza
Camila de Souza
GGG. Brian Golding
G. Brian Golding
LILucian Ilie
Lucian Ilie
Protein embeddings are the new main source of information about proteins, producing state-of-the-art solutions to many problems, including protein interaction prediction, a fundamental issue in proteomics. Understanding the embeddings and what causes the interactions is very important, as these models lack transparency due to their black-box nature. In the first study of its kind, we investigate the inner workings of these models using XAI (explainable AI) approaches. We perform extensive testing (3.3 TB of total data) involving nine of the best-known XAI methods on two problems: (i) the prediction of protein interaction sites using the current top method, Seq-InSite, and (ii) the production of protein embedding vectors using three methods, ProtBERT, ProtT5, and Ankh. The results are evaluated in terms of their ability to correlate with six basic amino acid properties—aromaticity, acidity/basicity, hydrophobicity, molecular mass, van der Waals volume, and dipole moment—as well as the propensity for interaction with other proteins, the impact of distant residues, and the infidelity scores of the XAI methods. The results are unexpected. Some XAI methods are much better than others at discovering essential information. Simple methods can be as good as advanced ones. Different protein embedding vectors can capture distinct properties, indicating significant room for improvement in embedding quality.