-The work is devoted to the analysis of observational data on the activity of laboratory animals. The activity cycles obtained by measuring sensors installed in the laboratory are considered. The software tool has been developed that provides processing, visualization and analysis of the obtained data.
The article considers the need to develop tools for
processing, analysis and visualization of data obtained by
observing laboratory rats and reading their activity on a
running wheel using a sensor. According to [
At the Department of Animal and Human Physiology of the Samara National Research University named after Academician S.P. Korolev many different experiments and observations are carried out to study the behavior of laboratory animals. However, for high-quality and convenient conclusion about the experiment there are not enough tools that would provide detailed analysis in addition to processing and visualization of the obtained data.
The software package developed by the authors was implemented in the Python language. This high-level programming language is designed to analyze various data. In addition, the Python contains a huge number of libraries required for simple and convenient implementation of software solutions that include visualization of various graphs.
Circadian rhythms are one of the most important
biological rhythms of living beings. They are manifested in
rhythmic changes in many behavioral, biochemical and
physiological parameters for a period close to twenty-four
hours [
The mammals circadian rhythm is controlled by the
suprachiasmatic nucleus of the hypothalamus, which
contains the main oscillator of the organism [8]. Researches
have shown that many different influencing factors cause
changes in circadian rhythms [
Research of the circadian biological rhythms is actual
problem of modern biology and medicine. The increased
interest to this problem is related to the facts that the
functional characteristics of organs and systems of the body
is impossible without their circadian physiological and
biochemical rhythms. Disruption of the rhythms leads to the
pathological conditions [
Observations of rats are made using a special sensor installed on each running wheel. Each one-eighth of a wheel's turn represents one conventional unit, called a bin. For every five seconds of observation, the computer records the number of bins that a particular rat makes during this period of time. Thus, the locomotor activity of a laboratory rat can be represented as a set of numbers.
It may notice that the rat has periods of decreasing activity when the rat is asleep, and periods of increasing when the rat is awake. These events are clearly described using the quantitative calculation of the partial revolutions of the running wheel. Based on the data obtained, it is possible to determine the period of locomotor activity of the rat. Thus, changes in the circadian rhythm are monitored over a long period of time, including before and after the influence of various factors.
The initial data received from the sensors is presented in the form of a table with six columns in a text document of the “.TXT” format (fig. 1).
Each column contains the following information: time since the start of observation, the number of bins made by five rats with a time interval of five seconds. For informational purposes, the name of each file is mainly the period of the experiment on rats indicating the beginning and end of observation with an accuracy of up to a minute.
Obviously, for a better analysis it is necessary to conduct long-term observations. However, this affects the number of rows in the resulting data document. This may result in a very long table. Therefore, for convenience, using the developed software tool, all rows in the table are grouped with a time interval (step) that the user sets himself (rows are initially grouped in 5 second increments). In this case, the bins of each rat in the table will display the number of partial revolutions of the running wheel in the required period of time. The data processor asks for the necessary interval at which the grouping will be performed. It also warns about errors and whether ungrouped lines remained due to the multiplicity of the total time to the entered step. Finally the user gets the resulting set of bins for each rat, as in the example in Figure 2.
In addition to the output of the processed data to the
console, it is possible to enter them into separate text
documents individually for each rat with a breakdown by the
number of days passed (Figures 3-4). The resulting files are
entered in a specially created directory with a unique name,
depending on the processing step and the name of the file
being processed. For this action to create a directory, the OS
library is used, which is an interface for portable use of the
operating system functionality [
An important feature of the developed software is the
visualization of the processed data. To implement this
functionality, the matplotlib package is used. Matplotlib is a
library for building various kinds of two-dimensional graphs,
histograms, diagrams in the Python with just a few lines of
code [
In the developed application, a pop-up window is implemented for viewing, editing and scaling of five graphs on the behavior of each rat based on the processed data with a user-selected step as shown in Figure 5.
A comparative analysis of the behavior of a laboratory rat
before and after exposure to any external facts consists in
determining changes in the following indicators: the period
of its circadian rhythm, acrophase and mesor, which describe
the motor activity of the rat [
Finding a period of circadian rhythm is a rather nontrivial task. The difficulty is as follows. The period is calculated from the usual sequence of numbers (bins for a certain period of time), the distribution of which is rather approximately described by a sinusoid. The authors offer a fairly simple way to determine the period. It is based on the logic of the distribution of numbers and on one of the methods for finding the period of a sinusoid by a graphical method.
This method describes the time intervals between the maximum achieved values of the number of bins (amplitudes) over time periods equal to 24 hours, that is, equal to the approximate value of the circadian rhythm in the laboratory animal. The feature of the software implementation of the approach is as follows. The user can enter the start time of the day for each laboratory animal (for the subsequent discarding of excess data obtained on the first day before a given point in time).
This procedure is necessary in order for the general locomotor activity of the test subject to be described by a cosine sine. This is necessary because the peak of activity in the rat is achieved at night. The algorithm for finding the biorhythm period is as follows.
The data is divided into equal time periods of 24 hours. At each interval, the amplitude is calculated. Then the time intervals between the nearest amplitudes are calculated, which are approximate values of the period of activity of the laboratory animal.
For a more accurate determination of the period, the arithmetic average of the obtained preliminary periods is calculated. According to the proposed method, the first value from the processed data can be taken as the first amplitude, since this is where the peak of activity is theoretically achieved. In addition, the first time period considered is usually 12 hours. A schematic representation of the algorithm is shown in Figure 6.
It should be noted that during the operation of the algorithm abnormal preliminary periods can be thrown.
For example, this happens when one amplitude was found close to the end of the considered segment, and the other amplitude was calculated at the beginning of the next segment. This means that the junction of the segments is at the peak of the locomotor activity of the rat. As a result, the algorithm receives a critically small period value due to the two resulting maxima.
As a rule, such an anomaly in the algorithm provokes the opposite anomaly. With this anomaly, the resulting sample of periods will contain a critically large gap between the amplitudes. This is because the first amplitude is calculated at the beginning of the interval, and the next at the end of the other. As a result, a huge period of circadian rhythm for the laboratory rat enters the algorithm. Two opposite anomalies when calculating the average period value from the sample cancel each other out. In this case, a value close to true is obtained.
The next characteristic – an acrophase is determined by the moments of time when the rat's activity amplitude is reached. The method of finding the acrophase of circadian rhythm was chosen to find the arithmetic mean value among the moments of time when the amplitudes were reached. In other words, for each amplitude, its deviation from midnight in real time was calculated, then the average value among the received deviations was determined, which was taken as the acrophase of the circadian rhythm of the rat in question.
The mesor is the value of the average level of a sinusoid
[
Next, cosinor analysis is carried out - the method for
approximating daily curves by harmonics (sinusoids) with a
certain period, based on the least squares method [
To construct a sinusoid (Figure 7), in addition to the period and acrophase, a mesor and amplitude are required.
The calculated function for the sinusoid is represented by the formula (1), where M is the mesor, А – amplitude, "φ" – acrophase, "τ" – period, е(t) – random component describing other factors not included in the function.
Y(t)=M+A* cos (2πt + φ) +e(t) τ (1)
Each of the obtained indicators is compared with the corresponding indicator before and after the influence of various influencing factors. Based on the comparison of indicators, the results and patterns of the influence of a factor on the behavior of a laboratory rat are formed.
To conduct a comprehensive analysis using the
developed software, we used data from observations of
laboratory rats obtained for the periods from November 24,
2009 to December 3, 2009 and from February 27, 2010 to
March 8, 2010 at the Department of human and animal
physiology at Samara University. Based on these data, the
Department conducted a research of the effect of insulin on
the circadian rhythm of arbitrary locomotor activity in rats.
The result of the reseacrh was the most pronounced shift of
locomotor activity in the direction of phase advance at the
time of injection of the medicine equal to ZT = 13 (analysis
of the acrophase shift revealed a significant advance of 5.48
± 1.98 hours), a slight change in the total daily activity [
The results of data visualization by the software package are shown in Figure 8. The data obtained as a result of the analysis, shown in Table 1, generally correspond to the previously obtained results of the research on the effect of insulin on the locomotor activity of rats. In the future, the authors plan to refine the algorithms and methods for finding the main characteristics. 101.5 160.5 113.5 162 42 89.5 153 61
The software tool designed to conduct a comprehensive analysis of the effect of various drugs on locomotor activity allows you to process rat observation data, present them in a visual form, and also determine the main characteristics of the circadian rhythm: activity period, acrophase, mesor. In addition, a function is constructed that describes the activity of laboratory animals in time. The software tool allows to analyze changes in the relevant characteristics before and after the influence of various influencing factors. The authors plan to expand the functionality of the developed software. In its final form, the software tool will allow more efficient and faster research of circadian biological rhythms and contribute to the emergence of new scientific discoveries.