Machine-learning model selection and parameter estimation from kinetic data of complex first-order reaction systems

Dealing with a system of first-order reactions is a recurrent issue in chemometrics, especially in the analysis of data obtained by spectroscopic methods applied on complex biological systems. We argue that global multiexponential fitting, the still common way to solve such problems, has serious weaknesses compared to contemporary methods of sparse modeling. Combining the advantages of group lasso and elastic net-the statistical methods proven to be very powerful in other areas-we created an optimization problem tunable from very sparse to very dense distribution over a large pre-defined grid of time constants, fitting both simulated and experimental multiwavelength spectroscopic data with high computational efficiency. We found that the optimal values of the tuning hyperparameters can be selected by a machine-learning algorithm based on a Bayesian optimization procedure, utilizing widely used or novel versions of cross-validation. The derived algorithm accurately recovered the true sparse kinetic parameters of an extremely complex simulated model of the bacteriorhodopsin photocycle, as well as the wide peak of hypothetical distributed kinetics in the presence of different noise levels. It also performed well in the analysis of the ultrafast experimental fluorescence kinetics data detected on the coenzyme FAD in a very wide logarithmic time window. We conclude that the primary application of the presented algorithms-implemented in available software-covers a wide area of studies on light-induced physical, chemical, and biological processes carried out with different spectroscopic methods. The demand for this kind of analysis is expected to soar due to the emerging ultrafast multidimensional infrared and electronic spectroscopic techniques that provide very large and complex datasets. In addition, simulations based on our methods could help in designing the technical parameters of future experiments for the verification of particular hypothetical models.

Electrostatic control of photoisomerization pathways in proteins

Rotation around a specific bond after photoexcitation is central to vision and emerging opportunities in optogenetics, super-resolution microscopy, and photoactive molecular devices. Competing roles for steric and electrostatic effects that govern bond-specific photoisomerization have been widely discussed, the latter originating from chromophore charge transfer upon excitation. We systematically altered the electrostatic properties of the green fluorescent protein chromophore in a photoswitchable variant, Dronpa2, using amber suppression to introduce electron-donating and electron-withdrawing groups to the phenolate ring. Through analysis of the absorption (color), fluorescence quantum yield, and energy barriers to ground- and excited-state isomerization, we evaluate the contributions of sterics and electrostatics quantitatively and demonstrate how electrostatic effects bias the pathway of chromophore photoisomerization, leading to a generalized framework to guide protein design.

Analysis of time-correlated single photon counting data: a comparative evaluation of deterministic and probabilistic approaches

We review various methods for analysing time-resolved fluorescence data acquired using the time-correlated single photon counting method in an attempt to evaluate their benefits and limitations. We have applied these methods to both experimental and simulated data. The relative merits of using deterministic approaches, such as the commonly used iterative reconvolution method, and probabilistic approaches, such as the smoothed exponential series method, the maximum entropy method and recently proposed basis pursuit denoising (compressed sensing) method, are outlined. In particular, we show the value of using multiple methods to arrive at the most appropriate choice of model. We show that the use of probabilistic analysis methods can indicate whether a discrete component or distribution analysis provides the better representation of the data.

Kinetics of Light-Induced Intramolecular Energy Transfer in Different Conformational States of NADH

When bound to a protein, the coenzyme NAD+/NADH typically exists in an extended conformation, while in aqueous solutions it can be characterized by an equilibrium of folded and unfolded structures. It was recognized long ago that in the folded conformation light absorption at the adenine ring initiates an effective energy transfer (ET) toward the nicotinamide group, but the mechanism of this process is still unexplored. Here we apply ultrafast transient absorption measurements on NADH combined with compartmental model analysis for following the kinetics of the ET. We find that the actual ET is extremely rapid (∼70 fs). The high rate can be well described by a Förster-type mechanism, promoted by both the special photophysical properties of adenine and the subnanometer inter-ring distance. The rapid ET creates a vibrationally hot excited state on nicotinamide, the vibrational and electronic relaxation of which is characterized by 1.7 and 650 ps, respectively.

Sample-pretreatment-free based high sensitive determination of aflatoxin M1in raw milk using a time-resolved fluorescent competitive immunochromatographic assay

A highly-sensitive TRFICA was developed to detect AFM₁in raw milk samples within 6 minutes without any sample pretreatments.