Research group of Dr. Mihaela Žigman

The levels, chemical modifications, and relative abundance of biomolecules that circulate in systemic bioliquids like human blood are a direct indicator of the physiological states of many organs within our body. Capitalizing on the broadband optics, ultrafast sources, and precision femtosecond-attosecond field-resolving metrologies of attoworld, we develop electric-field molecular fingerprinting (EMF) as a new analytical technique. Overcoming limitations of sampling depth and sensitivity inherent to existing omics technologies, EMF is capable of describing the entire ensemble of biomolecules, even in highly complex matrices like blood plasma. In strategic partnership with the Center for Molecular Fingerprinting (CMF), we work towards advancing EMF to an approach suited for high-throughput human populational screening enabling a robust novel route for health state monitoring and earliest possible disease detection, as described in our Just Cause

Due to pervasive connectivity at each level of regulation in our body (from genetic networks to cell-cell signaling to interactions with the environment), health as well as disease manifest as complex coinciding changes to all types of molecular classes – including proteins, lipids, nucleic acids, carbohydrates, metabolites. As EMF sensitively and robustly reports on all types of chemical bonds, it provides a molecular description of physiological phenotypes in their entirety. Measuring blood, whose chemical composition mirrors – to a significant extent – the cell functions in many tissues of our body, we develop EMF as a novel minimally-invasive tool for biomedicine.

Infrared vibrational spectroscopy analyzes electric fields originating from vibrations of the covalent bonds within molecules of the sample under investigation. During EMF, the entire molecular ensemble is probed within a single measurement, whereby the frequencies of laser pulse-induced vibrations vary depending on the exact atomic composition, structure, and strength of the chemical bonds in the individual molecules. Because infrared fingerprinting is inherently sensitive to all functional groups of organic compounds, it is well suited to probe biomolecules, which are often large and carry chemically complex modifications.

The EMF represents a highly characteristic, sensitive cross-molecular snapshot with many fine-grained features. We take advantage of machine learning algorithms to analyze patterns in these features and correlate these with person-specific medical, clinical, and lifestyle information. Eventually, EMF serves for in vitro detection of any physiological or pathological phenotype that leaves a molecular trace in the cell-free fraction of our blood.

• We exploit the potential of electric-field molecular spectroscopy to acquire molecular fingerprints of biological samples that contain highly complex, yet phenotype-specific mixtures of organic molecules. The ultrafast light sources and precision femtosecond field-resolving metrologies are being developed and continuously extended in our laboratories at attoworld.
  
  
• We probe biological samples in their native, hydrated, liquid phase (blood serum, plasma, exprimate urine, living cells, plant tissues, etc.). The approach operates in a label-free manner, thereby obviating the need for complex biochemical sample preparation. To make the practical workflow as effective as possible, we are combining electric-field spectroscopy with microfluidic sample delivery systems – to be using as little sample volume as possible, and rendering it as high-throughput as possible.
  
  
• Along with the development of the technology to measure samples, we have also established a cryogenic repository for the best long-term storage of biological materials, opening the possibility to measure and re-measure sample sets with ever higher sensitivity and specificity, as new generations of fingerprinting spectrometers are being developed.
  
  
• To facilitate the highest control over reproducibility and to eliminate possible confounding effects of sample preparation in real life outside the research lab (e.g. in medical clinics), we have worked out standard operating procedures along which we are, together with our clinical collaborating partners, running several clinical studies (for more information, please visit Laser4Life).
  
  
• Due to the complexity of patterns in the newly acquired electric-field spectra (the fingerprints), new ways of data processing and exploration are required, and we apply a diverse set of machine learning algorithms to build classifier models that facilitate phenotype detection.