Prophet 5.0 is an integrated software package designed to provide biomedical researchers with the data analysis tools essential for operating a research laboratory. Developed by BBN Technologies, a unit of GTE Internetworking, Prophet offers advanced statistical analysis features, mathematical modeling capabilities, and DNA analysis tools that are sometimes lacking even in "full-featured" molecular biology programs. Development is sponsored by the National Center for Research Resources of the National Institutes of Health. Prophet's integrated environment makes it an attractive all-in-one package for laboratory molecular biologists who appreciate the convenience and consistency of using a single tool for both numerical and sequence data analysis. Prophet 5.0 is available for Windows 95 and Windows NT as a free download from the BBN Web site, or may be ordered on CD-ROM for a small fee.

Prophet is a multifunction package allowing analysis of numerical data and genetic sequence data. The interface is user-friendly and intuitive. Those accustomed to the Windows environment will be quite familiar with Prophet's user interface. Upon launching, Prophet prompts users for information on the data to be analyzed. Users can choose to create a new sequence, table, or worksheet for numerical data. Once the appropriate data type is determined, users can easily enter the data or add a new amino acid or nucleotide sequence. In addition to the most important DNA analysis features, Prophet provides a number of built-in statistical tests, as well as a powerful nonlinear curve fitter.

Figure 1

Figure 2

The table interface of Prophet allows easy importing and exporting of data. The tables can also be reorganized, and subsets from the data table can be easily extracted (figure 1). Prophet's table interface is flexible enough to be used as a standalone spreadsheet-like program to store numerical data as well as text annotations. This allows a complete record of an experiment to be created in a single table (figure 2). Entire tables can be printed, or selected rows can be pooled into subsets for printing.

Figure 3

Figure 4

Prophet supports six different types of plots (figure 3). Once a graph is generated, it is fairly easy to manipulate the display by adding data and function curves to a graph, cutting and pasting different curves, and updating and annotating a graph (figure 4). Axes, legends, and titles can be manipulated to meet user preferences. One useful feature is that an "empty" X-Y series graph can be generated to act as a template for merging other graphs, enabling presentation of a comparative data analysis plot. Unfortunately, pie chart graphs are not supported in this version of Prophet.

Figure 5

Statistics is one of Prophet's most impressive features. A wealth of parametric and nonparametric tests are provided, as well as exploratory data analysis and curve-fitting functions. The statistical tests provided can be performed on data in a table or from a subset of selected data. The distribution tests include normal, uniform, chi-square, and goodness of fit, as well as t- and F- distributions. T-tests with paired or unpaired samples, and ANOVA (one way blocked, unblocked, and multiple) as well as a number of nonparametric tests, are built into Prophet. The nonparametric tests include one- and two-sample sign tests, one- and two-sample Wilcoxon signed rank tests, Mann-Whitney rank sum, Kruskal-Wallis, Friedman's, Ansari-Bradley, Cochran's Q, and McNemar's Q. ANCOVA, frequency and contingency categorical tests, and variance tests such as chi-square, Levene's, Bartlett's, and F-tests are included in Prophet, in addition to survival analysis, several exploratory data analysis methods, and regression tests. Built-in linear regression models (linear, multiple linear, and polynomial) are supplemented with an impressive list of nonlinear functions. In addition, data can be exported to SAS for further analysis. The output of Prophet's statistical tests is very clear and informative (figure 5), and some pages have direct Web links to more detailed information in BBN's StatGuide, a Web resource providing information on statistical tests as well as help on choosing and running them.

Figure 6

Figure 7

Once the Create New Sequence option is selected, DNA, RNA, or protein sequences can be imported, pasted, or manually entered into a sequence window. Prophet supports most sequence formats including GenBank, EMBL, and PIR. Each nucleotide and amino acid is color-coded (figure 6). There are options for automatically displaying the complementary strand of DNA sequences and for viewing the sequence in blocks of color, which allows quick visual identification of patterns. Multiple sequences can be included in a single file or can exist in separate files. A chemical composition tool easily calculates the composition and molecular weight of each sequence and gives additional information such as percent GC and the appearance of certain dinucleotide pairs for DNA sequences. Specific consensus sequence patterns can be identified with the pattern list, which includes sequence information for TATA, CAAT, and Pribnow boxes, among others. PCR primers can be effectively designed in Prophet, and this feature is very convenient and complete. Users can calculate the optimal areas in a sequence that the best PCR primer pairs can be designed, taking into consideration important information about the primers such as %GC, Tm, and optimal primer and salt concentrations in a PCR reaction (figure 7).

Figure 8

Similarly, Prophet facilitates the design of hybridization probes. A simple translation tool can translate a DNA sequence in all six reading frames. Tools for the identification of restriction enzymes and proteolytic enzymes for sequences are also part of the package. Multiple sequence alignments can be performed, and consensus sequence information produced, from multiple alignment data. Color coding and the option to view nucleotides and amino acids in blocks of different color makes this alignment feature a very powerful visual analysis tool (figure 8). One additional feature of Prophet is that a user can directly use NCBI's BLAST over the Web though a direct menu link. When this menu item is selected, Prophet automatically launches the default Web browser and accesses the appropriate site. The selected sequence is automatically converted into FASTA format and is placed in the clipboard, ready to be pasted into BLAST input form. Similarly, users can select Keyword Search and Retrieval from the Database menu to launch the default Web browser and link to the NCBI Entrez Web site to perform database searches for nucleotide and protein sequence information.

Figure 9

In addition to linear regression, Prophet boasts impressive nonlinear curve-fitting features using either simple Levenberg or Levenberg-Marquardt algorithms. Prophet includes built-in models for pharmacokinetic and pharmacodynamic analysis, as well as equations to analyze both direct and competition binding for one, two, and three binding sites, with parameters to take nonspecific binding into consideration. Dose-response data can be fitted with the built-in logistic model, and users can fit one, two, or three exponentials with or without constant offsets. In addition to these built-in models, users can enter their own model equations using a number of built-in mathematical functions. The data shown in figure 9 were fitted using a two-exponential equation entered into Prophet's Formula dialog box. Statistical results from nonlinear curve fitting include a number of residual plots, an ANOVA table, and goodness-of-fit data, as well as a table of the estimated parameters with relevant statistics such as standard deviation and 95% confidence intervals.

Prophet 5.0 is an integrated analysis program that can handle both numerical and sequence data as well as related annotations. Prophet's numerical, graphical, statistical, and sequence analysis features are well implemented, and the flexibility of its table structure provides a comprehensive package to collect, annotate, and analyze data. The sequence analysis package is useful for scientists that clone genes and probe for their function (PCR primers, hybridization, restriction enzyme site analysis), as well as for more complicated sequence analysis problems such as gene family creation and pattern recognition. Prophet offers features, including the ability to accomplish multiple sequence alignments, usually handled by separate "full-featured" software packages, without forcing users to sacrifice functionality. The seamless integration of Prophet with Web-based resources represents the leading edge of software development for personal computers.

Prophet 5.0 is available for Windows 95, Windows NT 3.51, or Windows NT 4.0. BBN recommends a 486 or better processor, 16 Mb (Windows 95) or 24 Mb (Windows NT) of RAM, and 20 Mb of disk space. Unix versions of Prophet 5.1 will be available in the near future for DEC Alpha, Sun SPARCstations running Solaris 2.5 or higher, and Silicon Graphics (including SGI Indigo) running IRIX 6.2 or higher.

Prophet 5.0 is available as a free download from BBN Technologies. The Prophet group can be reached by mail at BBN Technologies, 10 Moulton Street Cambridge, MA 02138; by phone at (800) 940-0103, or by email. Customers may purchase Prophet 5.0 on CD-ROM, which includes a set of printed manuals, for $40. Read more about Prophet's features at its Web site.