deletions | additions       diff --git a/Concepts.tex b/Concepts.tex  index 468dc1d..2fafd23 100644  --- a/Concepts.tex  +++ b/Concepts.tex 
...
\section{Architecture and Concepts}  \label{sec:concepts}  In this section, we introduce some general burst analysis  concepts and notations used in FRETBursts. \subsection{Photon Streams}  \label{sec:ph_streams} 
...
In single-spot measurements, all timestamps are stored in a single array. In multispot  measurements~\cite{Ingargiola_2013}, there are as many timestamps arrays  as excitation spots.  Each array contains timestamps from both donor (D) and acceptor (A) channels.  When alternating excitation lasers are used (ALEX measurements)~\cite{Lee_2005},   a further distinction between photons emitted during the D or A excitation periods can be made.   In FRETBursts, the corresponding sets of photons are called ``photon streams'' and are  specified with a \verb|Ph_sel| object  (\href{http://fretbursts.readthedocs.org/en/latest/ph_sel.html}{link}). 
...
\label{sec:bg_intro}  An estimation of the background rates is needed to both select a proper threshold for  burst search, and to  correct the raw burst counts by subtraction of background counts. The recorded stream of timestamps is the result of two processes: one characterized  by a high count rate, due to fluorescence photons of single molecules crossing the  excitation volume, and another characterized by a lower count rate, due to "background  counts" ``background  counts''  originating from detector dark counts, afterpulsing, out-of-focus molecules and sample scattering and/or impurities~\cite{Edman_1996,Gopich_2008}.  The signature of these two types of processes can be  observed in the inter-photon delays distribution (i.e. the waiting times  between two subsequent timestamps) as illustrated in figure~\ref{fig:bg_dist_all}(a).  The “tail” ``tail''  of the distribution (a straight line in semi-log scale) corresponds to exponentially-distributed time-delays, indicating that those counts are generated by a  Poisson process. At short  timescales, the distribution departs from the exponential due to the contribution  diff --git a/Concepts_3.tex b/Concepts_3.tex  index cffa49f..e468854 100644  --- a/Concepts_3.tex  +++ b/Concepts_3.tex 
...
FRETBursts allows performing the burst search on arbitrary selections of photons.  (see section~\ref{sec:ph_streams} for more information on photon stream definitions).  Additionally, Nir~\textit{et al.}~\cite{Nir_2006} proposed DCBS ('dual-channel (``dual-channel  burst search'), search''),  which can help mitigating artifacts due to photophysics effects such as blinking. During DCBS, a search is performed in parallel on two photon streams  and bursts are defined as periods during which both photon streams exhibit a rate higher than  the threshold, implementing the equivalent of an AND logic operation. 
...
are (1) all photons during donor excitation (\verb|Ph_sel(Dex='DAem')|) and  (2) acceptor channel photons during acceptor excitation (\verb|Ph_sel(Aex='Aem')|).  In FRETBursts, the user can choose arbitrary photon streams as input, an in general  this kind of search is called a 'AND-gate ``AND-gate  burst search'. search''.  After burst search, it is necessary to select  bursts, for instance by specifying a minimum number of photons (or burst size). In the most 
...
the burst size selection threshold is weighting bursts proportionally to their size   so that the bursts with largest sizes will have the largest weights.  Using size as weights (instead of any other monotonically increasing function  of size) can be justified noticing that the variance of bursts PR ($E_i$) proximity ratio (PR)  is inversely proportional to the burst size (see SI~\ref{sec:burstweights_theory} for details).   In general, a weighting scheme is used for building efficient estimators for a population  parameter (e.g. the population FRET efficiency  $E_p$). But, it can also be used to build weighted histograms or Kernel Density Estimation (KDE) plots which emphasize FRET subpopulations peaks   without excluding small size bursts.  Traditionally, for optimal results when not using weights, the   diff --git a/figures/ph_delays_distrib_all/caption.tex b/figures/ph_delays_distrib_all/caption.tex  index f3bd7a1..c0a3a75 100644  --- a/figures/ph_delays_distrib_all/caption.tex  +++ b/figures/ph_delays_distrib_all/caption.tex 
...
(\textit{Panel a}) An example of inter-photon delays distribution (\textit{red dots}) and an exponential fit  of the tail of the distribution (\textit{black line}).  (\textit{Panel b}) Inter-photon delays distribution and exponential fit for different photon streams as obtained with \texttt{dplot(d, hist\_bg)}. The \textit{dots} represent the experimental histogram for the different photon streams. The \textit{solid lines} represent the corresponding exponential fit of the tail of the distributions. The legend shows abbreviations of the photon streams  and the fittedrate  background rate. rates.