= Ilya J. Finkelstein
{c}
\b[Ilya J. Finkelstein] is a researcher in molecular biophysics whose work includes ultrafast molecular measurements and methods for investigating protein–DNA interactions. His own curriculum vitae identifies him by this full byline, records his Stanford chemistry doctorate and describes later single-molecule and parallel biophysical measurements. The initial J. is not expanded here.<Ilya J. Finkelstein curriculum vitae and research profile>[CV, heading, Education and Five Significant Publications]{ref}
<Ilya J. Finkelstein> coauthored the 2007 <2D IR center-line slope> paper with <Kyungwon Kwak>, <Sungnam Park> and <Michael D. Fayer>.<Frequency-frequency correlation functions and apodization in two-dimensional infrared vibrational echo spectroscopy: A new approach by Kyungwon Kwak and others>[author byline]{ref} The paper addresses a practical <inverse problem>: extracting molecular frequency-memory information from measured spectra, rather than identifying that information with any convenient peak width.
His later research profile describes methods that resolve interactions involving individual molecules and many distinct molecular substrates.<Ilya J. Finkelstein curriculum vitae and research profile>[CV, Five Significant Publications; laboratory Members, Ilya Finkelstein research description]{ref} An original counting model explains one benefit of resolving separate events. If $N$ independent, equally observable sites each bind a probe with probability $p$, the bound count $B$ has
$$
\mathbb E[B]=Np,\qquad \operatorname{Var}(B)=Np(1-p).
$$
Consequently $\widehat p=B/N$ has <standard deviation> $\sqrt{p(1-p)/N}$. At stipulated $N=100$ and $p=1/2$, that scatter is 0.05. A single observed fraction 0.55 is therefore not by itself evidence for a different binding mechanism.
The assumptions are substantive. Sites may interact, molecules may be missed, and a bright object may contain more than one labeled molecule. <Background subtraction>, detection efficiency and an appropriate <measurement model> determine whether image counts represent binding events. Averaging more frames does not remove a persistent detection bias.
The common methodological theme is quantitative access to otherwise hidden dynamics: controlled spectroscopy supplies frequency correlations, while molecule-resolved assays can separate populations hidden by an ensemble mean. Neither method replaces controls or a model linking the signal to the molecular quantity of interest.
\Image[https://finkelsteinlab.org/assets/images/members/Ilya-at-bench.jpg]
{title=<Ilya J. Finkelstein>, photograph linked from his named laboratory curriculum vitae}
Back to article page
