Genomic analysis gallery

Complex genomic analysis for research and precision medicine requires numerous methods for robust statistical validation. However, to fully digest the raw data and present results that are both actionable and understandable, data-tailored presentation is critical. Presentations must by accurate, self-contained, and simplified. Often, the best data may be complex, multi-dimensional, or contain novel information that does not conform to standardised presentations. This gallery contains some examples of both simple and complex visualisations that we have previously published. This page is under live development - text descriptions and links to primary literature will be available shortly.

⚡ Open live data
Presenting functional evidence for genomic determinants of disease that are succinct, interactive, and aesthetic is a difficult task and often overlooked. This project aims to make a public resource to empower clinically actionable genetics.

⚡ Open live data
A large number of evidence sources are used for bioinformatic analysis and annotation. However, these tools are inaccessible to non-bioinformaticians. This browser allows the public to directly link to the same evidence sources interactively.

⚡ Open live data
Edward Tufte calls Minard's graphic of Napoleon in Russia one of the "best statistical drawings ever created". This version brings the data to life. Built as a test for a modular dataviz system used in a genetic analysis browser.

⚡ Open live data
Data usage tracking - Deep Space Genomics©. Long-term projects re-use data, often dormant for long periods. We built systems that are used to present summarised analytics, conducive to multipurpose usage and error prevention by automated data tracing. Scans project code to find data sources and produces live interactive reports, without manual curation. github.com snippet example

⚡ Open live data
Somatic variant risk - Deep Space Genomics©. Individual events among crew members over ten decades. Specific critical events are shown by black circles. github.com

⚡ Open live data
Somatic variant risk - Deep Space Genomics©. Cumulative risk over ten decades. General thresholds for pre-emptive and critical care are shown with hlines. Specific critical events are shown by black circles. github.com

⚡ Open live data
Fictional data for real-time financial statistics in property investment planning. Monthly change with fixed return and dynamic expenses. github.com

⚡ Open live data
Fictional data for real-time financial statistics in property investment planning. Daily change with multiparty investors. github.com
Protein network analysis for patients with shared phenotypes.
A splice site variant causes dominant autoinflammatory disease. (A) The pedigree includes Sanger sequences showing inheritance pattern, (B) cDNA sequence shows the outcome on transcript, (c) the effect on protein haploinsufficiency, (D) coding sequence illustration. doi:10.3389/fimmu.2018.01527
Variants in an autoinflammatory disease-causing protein are predicted based on population genetics data and illustrated against the multiple, critical functional domains in the protein. Residues that are highly conserved and prone to de novo mutation are most likely to require clinical intervention. doi:10.3389/fimmu.2018.01527
Combined reporting of raw data and the complete analysis for population genetics studies. This report contains all data and analysis scripts to reproduce the results and figures as illustrated in this plan.
Visualise novel analysis methods; predicted amino acid variation in critical immune proteins. doi:10.1007/s10875-019-00670-z
Variant pathogenicity compared probability of variant in patients. doi:10.1007/s10875-019-00670-z
Variants that are predicted to be the most clinically-relevant are compared to the proteins structure (RAG protein complex binding DNA); likelihood of variants presenting as disease in humans; blue - not likely, yellow - mild, orange - moderate, red - most likely. DNA are shown in green undergoing recombination during immune system development. doi:10.1007/s10875-019-00670-z
Novel clinical genetics diagnostics for rare disease. doi:10.1002/art.41531
Complex large-scale data storage plans.
Complex large-scale data analysis summaries.
Consulting with examples for genomic analysis project planning.
Protein structure of SARS-CoV-2 S protein structure with variants-of-concern used for vaccine resistance planning. sarscov2variants.com
Complex clinical history of patients showing both the treatments and disease features. Patients with the same rare disease often have a range of unusual phenotypic features. This information can be difficult to compare by tables or written text. A visualisation can assist the reader in digesting subtle differences and commonalities. doi:10.1182/blood.2020005844 pdf
Ensuring that false positives do not affect mutation probability prediction based on evolutionarily constrain genomic regions. doi:10.1007/s10875-019-00670-z
Germline antibody gene locus. Custom illustrations based on several figures from Janeway’s Immunobiology, Garland Science.
Germline T cell receptor gene locus. Custom illustrations based on several figures from Janeway's Immunobiology, Garland Science.
Lymphocyte development. Progenitor cells must undergo several maturation processes during lymphocyte development. A simplified illustration of diversification of the T and B cell repertoire is shown occurring via the thymus and bone marrow, respectively.
RAG1 PHRED-scaled CADD score versus GnomAD conservation rate and MRF score, predicting clinical presentation of variants. doi:10.1007/s10875-019-00670-z
Communicating details of rare disease cohorts during public presentation. National-scale data for severe immune disorders.
Crystal structure of DNA bound RAG complex, based on 3jbw.pdb. 10.2210/pdb3JBW/pdb
All known disease-causing variants of RAG-deficiency mapped to protein domains (blue dots) and newly identified variants (red dots). 10.1016/j.jaci.2018.02.007
Translating complex gene recombination processes of lymphocyte development for clinical immunologists. Adaption based on De Villartay et al. Nature Reviews Immunology (2003).
DNA recombination complex formation. Illustration based on work from The elements of immunology Pearson 2009 and Janeway's immunobiology Garland Science, 2016.
The architecture of genetic factors in lymphoma patients with germline and somatic causal variants.
Variant frequency and consequence in population genetics. Data from 150K genomes to characterise one Mendelian disease gene.
GWAS analysis of disease-causing variants for immunological phenotypes.
Several databases of known pathogenic human variants were used to validate a new variant prediction score. 36% of all pathogenic variants were predicted as the top candidate with this method, out performing Phred CADD for rare causal variants and pre-empting clinically relevant variants.
Genomics of rare disease as part of national-scale genomics services.
Genomic analysis of the first complete germline loss of TET2 reported in humans. Validated by accredited clinical genetics and extensive functional work. doi:10.1182/blood.2020005844
I developed a new method for very cheap whole-genome methylation analysis by modifying tools that are available from NEB and Agilent. This image was used as part of the protocol for a simplified overview.
Genomic discovery in rare disease often requires extensive functional validation. Here we express pathogenic variants from patients in vitro for validation. doi:10.1016/j.jaci.2018.02.007 and doi:10.1182/blood-2018-07-866939
Genomic discovery in rare disease often requires extensive functional validation. Here we express pathogenic variants from patients in vitro for validation. doi:10.1016/j.jaci.2018.02.007 and doi:10.1182/blood-2018-07-866939
Analysis in families can require contacting individuals worldwide, accurately gathering and interpreting data from patients, clinicians, and clinical geneticists - adhering to ethical and legal requirements and uncovering pertinent details for the patient.
Genetic discoveries can have serious implications that should be reported with care and consideration, as illustrated here for a family who were all heterozygous carriers of a variant that was lethal for one child.
Having the tools and knowledge to rapidly access large-scale reference datasets like gnomAD can vastly improve decision making during statistical analysis.
Understand both the known and unknown possibilities for real-world applications allows practical study design. Here, newly discovered clinically-relevant candidate variants are compared to those which have been tested in vivo and in vitro. doi:10.1007/s10875-019-00670-z
Description All known disease-causing variants of RAG-deficiency mapped to protein domains (blue and red dots). The orange/red and blue/pink bars show population based statistics of allele frequencies.
Statistical analysis of rare severe disease allowed prediction of pathogenic variants that have and will present clinically (left). A vast improvement over population based statistics prediction as used by many commercial and clinical labs (right).
Two genes where a multi step process (i-iii) illustrates the improvement for detecting causal variants. The small list of candidate variants allowed for functional validation before clinical presentation requiring urgent care. These would otherwise be defined as variants of unknown significance.
A synthetic design allowing for in vitro DNA recombination testing. doi:10.1182/blood-2018-07-866939
Keeping up to date with both private and public genomics programs worldwide for data access, planning, and collaboration.
Keeping up to date with both private and public genomics programs worldwide for data access, planning, and collaboration.
Logo image

Maintained by Dylan Lawless. Scientist at EPFL. Resume and Google scholar. RSS feed