Search the web
Welcome, Guest
[Sign Out, My Account]

Quotes & Info
Enter Symbol(s):
e.g. YHOO, ^DJI
Symbol Lookup | Financial Search
SLP > SEC Filings for SLP > Form 10-Q on 8-Jul-2014All Recent SEC Filings

Show all filings for SIMULATIONS PLUS INC



Quarterly Report

Item 2. Management's Discussion and Analysis or Plan of Operations

Forward-Looking Statements

This document and the documents incorporated in this document by reference contain forward-looking statements that are subject to risks and uncertainties. All statements other than statements of historical fact contained in this document and the materials accompanying this document are forward-looking statements.

The forward-looking statements are based on the beliefs of our management, as well as assumptions made by and information currently available to our management. Frequently, but not always, forward-looking statements are identified by the use of the future tense and by words such as "believes," expects," "anticipates," "intends," "will," "may," "could," "would," "projects," "continues," "estimates" or similar expressions. Forward-looking statements are not guarantees of future performance and actual results could differ materially from those indicated by the forward-looking statements. Forward-looking statements involve known and unknown risks, uncertainties, and other factors that may cause our or our industry's actual results, levels of activity, performance or achievements to be materially different from any future results, levels of activity, performance or achievements expressed or implied by the forward-looking statements.

The forward-looking statements contained or incorporated by reference in this document are forward-looking statements within the meaning of Section 27A of the Securities Act of 1933, as amended ("Securities Act") and Section 21E of the Securities Exchange Act of 1934, as amended ("Exchange Act") and are subject to the safe harbor created by the Private Securities Litigation Reform Act of 1995. These statements include declarations regarding our plans, intentions, beliefs or current expectations.

Among the important factors that could cause actual results to differ materially from those indicated by forward-looking statements are the risks and uncertainties described under "Risk Factors" in our Annual Report and elsewhere in this document and in our other filings with the SEC.

Forward-looking statements are expressly qualified in their entirety by this cautionary statement. The forward-looking statements included in this document are made as of the date of this document and we do not undertake any obligation to update forward-looking statements to reflect new information, subsequent events or otherwise.



Simulations Plus, Inc., incorporated in 1996, develops and produces software for use in pharmaceutical research and for education, as well as provides contract research services to the pharmaceutical industry. Two recent in-house developments take advantage of theartificial neural network ensemble modeling engine in our ADMET Predictor™ software for new markets. AEROModeler™ predicts missile aerodynamic force and moment coefficients and MRIModeler™ classifies patients as either likely to experience some form of autism or not. These are discussed further below under ADMET Predictor.

We currently offer five software products for pharmaceutical research: ADMET Predictor™, MedChem Designer™, MedChem Studio™, DDDPlus™, and GastroPlus™. We call the combination of ADMET Predictor, MedChem Studio, and MedChem Designer our ADMET Design Suite™.

ADMET Predictor™

ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) Predictor is a computer program that takes molecular structures as inputs and predicts about 145 different properties for them at the rate of over 100,000 compounds per hour on a fast laptop computer. This capability allows chemists to get estimates for a large number of important properties without the need to synthesize and test the molecules. ADMET Predictor has been consistently top-ranked for predictive accuracy in peer-reviewed, independent comparison studies, while generating its results at a very high throughput rate. Although the state-of-the-art of this type of software does not enable finding the best molecule in a series, it does allow identifying molecules that are highly likely to fail as potential drug candidates (the worst molecules, which is usually the majority of a chemical library) before synthesizing and testing them. Thus, millions of "virtual" compounds can be created and screened in a day, compared to potentially months or years of work to actually synthesize and test a much smaller number of actual compounds.

The ADMET Modeler™ subprogram that is integrated into ADMET Predictor enables scientists to use their own experimental data to quickly create high-quality, proprietary predictive models using the same powerful modeling methods we use to build our top-ranked property predictions. Pharmaceutical companies expend substantial time and money conducting a wide variety of experiments on new molecules each year, resulting in large databases of experimental data. Using this proprietary data to build predictive models can provide a second return on their investment; however, model building has traditionally been a difficult and tedious activity performed by specialists. The automation in ADMET Modeler makes it easy for a scientist to create very powerful models with a minimum of training.

We are now examining a very different application of this modeling engine - building predictive models for missile aerodynamic force and moment coefficients as a function of missile geometry, Mach number, and angle of attack. This problem was identified by the Aerospace Engineering department at Auburn University, and working with them, we have done some preliminary testing of the modeling engine in ADMET Modeler for this type of problem. Results have been very encouraging, and we believe there are government agencies and industrial aerospace companies that will find such a capability to be highly useful. We have developed a prototype AEROModeler™ program to test this concept and to use as a demonstrator for proposal efforts to potential funding agencies. Our proposed joint scientific poster on this subject with Auburn University's Aerospace Engineering Department was selected for presentation at the NSMMS/CRASTE (National Space and Missile Material Symposium/Commercial and Government Responsive Access to Space Technology Exchange) conference in Huntsville, Alabama in June 2014, and received significantattention and positive feedback from both government agencies and aerospace contractors, not only for aerodynamic coefficient predictions, but for several other problems of interest to the industry.

We have also begun a preliminary investigation of applying this powerful modeling engine to the analysis of MRI (magnetic resonance imaging) data in cooperation with the MRI facility at Auburn University. This state-of-the-art facility has two MRIs - one a 3-Tesla machine and one a very powerful 7-Tesla machine, both built in the last few years. We are examining data from a series of subjects in four groups: healthy, PDD (Pervasive Developmental Disorder), ADHD (Attention Deficit Hyperactivity Disorder), and Asperger's Syndrome, to determine whether we can discriminate between the groups from the MRI imaging data. The amount of data is massive ("big data"), requiring us to modify our code to handle much larger data arrays than our previous applications have required. Our current goal is to show the potential of our modeling technology to provide useful classification of a patient into one of the four groups based only on MRI imaging, so that we could go to various agencies (such as the National Institutes of Health) to obtain funding to develop a commercial product. Preliminary results exceed published classification accuracies from other groups by a substantial margin. We will be presenting a scientific poster at the Fourth Biennial Conference on Resting State/Brain Connectivity hosted by the Massachusetts Institute of Technology (MIT) in Cambridge, MA on September 10-13, 2014. We believe our artificial neural network ensemble modeling engine has wide-ranging applications and we intend to pursue funding to develop customized tools based on the engine for a number of potential applications.

During this reporting period, we released version 7.0 of ADMET Predictor. This new version incorporates a powerful new model for predicting ionization constants (pKa's), developed in a collaboration with Bayer AG that enabled us to more than double the size of our data set from about 16,000 pKa values to more than 35,000, and to expand the chemical space it covers to include a larger number of molecules more like those of interest to the pharmaceutical industry today. We believe the resulting improvement in pKa prediction puts our already best-in-class model well in front of any competitor. Predicting ionization is critical to predicting most other properties, so all of our models (approximately 144) were retrained based on this new capability for version 7.0. We are currently finalizing version 7.1, which incorporates several new capabilities that we expect to release during our 4th fiscal quarter.

MedChem Designer™

MedChem Designer was launched in 2011. It was initially a molecule drawing program, or "sketcher", but now has capabilities exceeding those of other molecule drawing programs because of its integration with both MedChem Studio and ADMET Predictor. We provide MedChem Designer for free because we believe that in the long run it will help to increase demand for ADMET Predictor and MedChem Studio, and because most other existing molecule drawing programs are also free. Our free version includes a small set of ADMET Predictor property predictions, allowing the chemist to modify molecular structures and then see a few key properties very quickly. The chemist also sees that with a paid ADMET Predictor license, the entire 140+ predictions would be available.

When coupled with a license for ADMET Predictor, MedChem Designer becomes a de novo design tool for medicinal chemists. With it, they can draw one or more molecular structures, then click on the ADMET Predictor icon and have over 140 properties for each structure calculated in seconds, including our proprietary ADMET Risk™ index. Scientists can also click on an icon to generate the likely metabolites of a molecule and then predict all of the properties of those metabolites from ADMET Predictor, including their ADMET Risk scores. This is important because a metabolite of a molecule can be harmful even though the parent molecule is not.

ADMET Risk provides a single number that tells the chemist how many default threshold values for 24 predicted properties were crossed (or violated) by each structure. The rules can be modified and new rules added by the user to include any desired rule set based on any combination of calculated descriptors, predicted properties, and user inputs. Thus, in a single number, the chemist can instantly compare the effects of different structural changes in many dimensions. As chemists attempt to modify structures to improve one property, they often cause others to become unacceptable. Without ADMET Risk, the chemist would have to individually examine many key properties for each new molecule (and its metabolites) to check whether any of them became unacceptable as a result of changing the structure.

During this reporting period, we released version 3.0 of MedChem Designer, which added the ability to capture the image of a molecular structure from a variety of publication files with a new snapshot tool, and then have the program automatically convert the graphic image into any of several computer-based chemical structure files. Converting from lines and letters on the screen to an exact chemical representation of the molecule (Optical Structure Recognition, or OSR) is a complex task. Although a few OSR programs are in existence, we are not aware of any that can accurately convert as many varieties of images to chemical representation as the OSR tool within MedChem Designer. Such a capability allows chemists to quickly capture molecular structures from the scientific literature to use for various purposes, including in our simulation and modeling software.

MedChem Studio™

While MedChem Designer can be used to refine a small number of molecules, MedChem Studio can be used to create and screen (with ADMET Predictor) a very large number of molecules down to a few promising lead candidates. MedChem Studio has features that enable it to generate new molecular structures using a variety of de novo design methods. Coupled with ADMET Predictor and MedChem Designer, we believe the programs provide an unmatched capability for chemists to search through large libraries of compounds that have undergone high-throughput screening experiments to find the most promising classes (groups of molecules with a large part of their structures the same) and molecules that are active against a particular target. In addition, MedChem Studio can take an interesting (but not acceptable) molecule and, using a variety of design algorithms, very quickly generate many thousands to millions of high quality analogs (similar new molecules). These molecules can then be screened using ADMET Predictor to find molecules that are both active against the target as well as acceptable in a variety of ADMET properties.

MedChem Studio version 4.0 was released during the current reporting period.

NCE Projects

During late 2012, based on our strong belief in the exceptional capabilities of our ADMET Design Suite (MedChem Studio/MedChem Designer/ADMET Predictor), we initiated a new molecule (NCE, or New Chemical Entity) design project. After considering various targets, we selected the malaria parasite Plasmodium falciparum, both because of the unmet need for a very low-cost cure, and because we believed that external funding opportunities might exist if we were successful in generating high-quality lead compounds using our software. Our goal was to demonstrate how well the ADMET Design Suite worked to generate new lead molecules in a fraction of the time and cost normally required in the pharmaceutical industry. We completed the design process in September 2012 and we announced that we had requested quotations from chemical synthesis companies for the cost and time to make a small set of molecules. Five molecules of our own design and two precursors (almost the final designed structures, but a step away in synthesis) were synthesized and tested for inhibition of the parasite at the University of California at Riverside. We were hoping that at least one would show inhibition of the growth cycle of the parasite.

We were excited to learn that every molecule showed activity against the parasite at less than micromolar concentrations, with two showing activity at less than 100 nanomolar concentration (high potency) against the drug-sensitive strain of the parasite. They were then tested against the newer drug-resistant strain of the parasite, and again potency was observed, with two molecules showing nanomolar activity. We believe this exercise - a software company using its own products to design novel molecules and have them synthesized and tested
- is unprecedented. New software license sales resulting from presenting our results have already more than recovered our investment.

During previous reporting periods, we announced that we had completed the design of a number of new molecules for a different target - the cyclo-oxygenase-2 (COX-2) enzyme that is the target for Celebrex®. Celebrex is the only COX-2 inhibitor remaining on the market, after the withdrawal of other approved drugs (such as Vioxx®) due to cardiac toxicity. It appears from the scientific research that was conducted after the withdrawal of other COX-2 inhibitors from the market that it is important to inhibit both COX-2 and COX-1 at a certain ratio in order to provide the benefits of COX-2 inhibition without the cardiotoxicity risk that has been associated with inhibiting COX-2 alone. We designed our new molecules based on activity models for both COX-2 and COX-1 built from public data, with the goal of providing an acceptable ratio of COX-2 to COX-1 inhibition. This is more challenging than designing for a single target, as we did for the earlier malaria NCE project. Results reported during this reporting period showed that we were once again very successful, with all four molecules that were synthesized inhibiting both the COX-2 and COX-1 enzymes, and one of them providing the desired characteristic of higher affinity for COX-2 than COX-1. We believe this is a remarkable and unique achievement among software companies, and clearly demonstrates that our ADMET Design Suite can save considerable time and money in developing new lead compounds for particular targets.


DDDPlus simulates in vitro laboratory experiments used to measure the rate of dissolution of the drug and, if desired, the additives (excipients) contained in tablets and capsules under a variety of experimental conditions. This software program is used by formulation scientists in industry and the U.S. Food and Drug Administration (FDA) to (1) understand the physical mechanisms affecting the dissolution rate for various formulations, (2) reduce the number of cut-and-try attempts to design new drug formulations, and (3) to design in vitro dissolution experiments to better mimic in vivo conditions.


Our flagship product and largest source of revenues is GastroPlus. GastroPlus simulates the absorption, pharmacokinetics, and pharmacodynamics of drugs administered to humans and animals, and is currently in widespread use at pharmaceutical companies, the FDA, the U.S. National Institutes of Health (NIH), and other government agencies in the U.S. and other countries. Because of the widespread use of GastroPlus, we were the only non-European company invited to join the European Innovative Medicines Initiative (IMI) program for Oral Bioavailability Tools ("OrBiTo"). OrBiTo is an international collaboration among 27 industry, academic, and government organizations working in the area of oral absorption of pharmaceutical products. Because we are outside of Europe, our participation in this project is at our own expense, while other members are compensated for their work; however, we are a full member with access to all of the data and discussions of all other members. We believe participation in this initiative enables us to benefit from and to contribute to advancing the prediction of human oral absorption from preclinical data, and ensures that we are in front of the audience of member pharmaceutical companies and regulatory agencies.

Version 8.5 of GastroPlus was released during the previous reporting period, adding a number of important new capabilities requested by customers as well as improvements we have identified in-house, including:

A new model for precipitation based on classical nucleation theory

Infant physiologies, including for babies born as much as 16 weeks premature

A unique method for using transporter data from preclinical experiments to predict transporter effects in human and other animals

A number of additional expression levels of enzymes and transporters in human and animal physiologies

An interim release (8.6) is planned for the 4th fiscal quarter (June 1 - August 31) to enable certain customers to take advantage of a new physiological model for minipig, which has become a more frequently used animal species in preclinical development, and to add the ability to simulate populations within the Drug-drug interaction Module.

The next major release, version 9.0, is already well along in development. This version will add the ability to simulate dermal (through the skin) drug absorption from patches, creams, and ointments. This capability was developed through a funded collaboration with a top-5 pharmaceutical company, and is already in use at the customer's sites at this time. A number of other improvements will be included in version 9.0 that will be announced with the release of the product.


MembranePlus is a new product that has been under development for a number of years, but was put on hold for several years due to other priorities. It was revived in the past year and is now nearing commercial release. Like DDDPlus, MembranePlus simulates laboratory experiments, but in this case, the experiments are for measuring permeability of drug-like molecules through various membranes, including several different cell cultures (Caco-2, MDCK) as well as artificially formulated membranes (PAMPA). The value of such a simulation results from the fact that when the permeabilities of the same molecules are measured in different laboratories, results are often strikingly different. These differences are caused by a complex interplay of factors in how the experiment was set up and run. MembranePlus simulates these experiments with their specific experimental details, and this enables the scientist to better interpret how results from specific experimental protocols can be used to predict permeability in human and animals, which is the ultimate goal. MembranePlus is unique and our customers have expressed significant interest in the new capability.

Priorities for developing material for our May training workshops as well as numerous conferences and on-site customer trainings required us to delay development of MembranePlus. We now plan to release version 1.0 of MembranePlus in the fourth fiscal quarter.

Contract Research and Consulting Services

Our expertise in oral absorption and pharmacokinetics is evidenced by the fact that our staff members have been speakers or presenters at over 150 scientific meetings worldwide in the past four years. We frequently conduct contracted studies for large customers (including the largest five pharmaceutical companies) who have particularly difficult problems and who recognize our expertise in solving them, as well as for smaller customers who prefer to have studies run by our scientists rather than to license our software and train someone to use it. The demand for our consulting services has been steadily increasing, and we have expanded our Simulations Studies team to meet the increased workload. Long-term collaborations and shorter-term consulting contracts serve both to expand and showcase our technologies, and to build and strengthen customer relationships.

During the current reporting period, we continued to work on our 5-year Research Collaboration Agreement (RCA) with the Center for Food Safety and Applied Nutrition (CFSAN) of the FDA. FDA scientists and our scientists are using ADMET Predictor/Modeler to build predictive models for likely toxicities of food additives and contaminants. During the first part of this collaboration, we analyzed FDA databases and worked with FDA scientists to ensure that the FDA data to be used for building new predictive models is as accurate as we can reasonably make it. Both FDA scientists and our scientists are building a series of models to classify new compounds as toxic or nontoxic from FDA datasets. Included early on in this effort was a special modification to ADMET Predictor to allow the user to set a minimum value for specificity or sensitivity when building a model, and this is now a standard part of the program available to all users. Sensitivity refers to how well a model identifies toxic (or any other property) compounds. A model that determined all compounds are toxic would have 100% sensitivity, because all toxic compounds would be labeled as such; however, all nontoxic compounds would also be labeled toxic. Specificity refers to how well a model distinguishes between toxic and nontoxic compounds. Increasing one usually results in decreasing the other. Depending on the purpose of the model, some scientists will prefer to train models that emphasize one statistic over the other.

Also during this reporting period, we began another five-year RCA, this time with the Office of Generic Drugs (OGD) within the FDA. This RCA is directed toward the FDA's evaluation of mechanistic IVIVCs (in vitro-in vivo correlations), an approach to determine whether mechanistic absorption modeling (MAM) correlates laboratory (in vitro) dissolution experiments with the in vivo behavior of a dosage form better than traditional empirical methods. We have proposed this method for about 15 years and believe in it, so we are pleased to see the FDA giving it serious consideration with this RCA.


Our business strategy is to do the things we need to do to promote growth both organically (by expanding our current products and services through in-house efforts) and by acquisition. We believe in the "Built to Last" approach - that the fundamental science and technologies that underlie our business units are the keys both to improving our existing products and to expanding the product line with new products that meet our various customers' needs.

With our significant cash reserves, seeking suitable acquisitions is a priority. Because we had been unable to identify suitable acquisitions and our cash continued to accumulate, the board of directors declared a $0.05 per share per quarter cash dividend that began in February 2012 and was paid in May, August, and November 2012. The board declared an accelerated cash dividend consisting of the February, 2013 dividend of $0.05 per share per quarter plus $0.03 per share from each of the expected May, August, and November 2013 dividends of $0.05 per share per quarter for a total of $0.14 per share, which was distributed on December 28, 2012, in order to provide our shareholders with the income tax benefits from lower capital gains rates in 2012 over 2013. We declared a $.04 per share dividend in November 2013 and $.05 per share dividends in February and May 2014. We anticipate the dividend to continue to be $0.05 per share per quarter, however there can be no assurances that such dividends will be distributed, or if so, whether the amounts will be more, less, or the same as expected. The Board of Directors must approve each quarterly dividend distribution and may decide to increase, decrease, or eliminate dividend distributions at any time.

Results of Operations

Comparison of Three Months Ended May 31, 2014 and 2013.

The following table sets forth our condensed statements of operations (in
thousands) and the percentages that such items bear to net sales (because of
rounding, numbers may not foot):

                                                  Three Months Ended
                                           05/31/14                05/31/13
Net sales                             $ 3,741       100.0 %   $ 3,095       100.0 %
Cost of sales                             228         6.1         452        14.6
Gross profit                            3,513        93.9       2,643        85.4
Selling, general and administrative     1,204        32.2         904        29.2
Research and development                  235         6.3         206         6.7
Total operating expenses                1,439        38.5       1,110        35.9
Income from operations                  2,074        55.4       1,533        49.5
Other income                               15         0.4          13         0.4
Income from operations before taxes     2,089        55.9       1,546        49.9
(Provision for) income taxes             (781 )     (20.9 )      (553 )     (17.8 )
Net income                            $ 1,308        35.0 %   $   993        32.1 %

Net Sales

Net sales Increased $646,000, or 20.9%, to $3,741,000 in the third quarter of Fiscal Year 2014 ("3QFY14") from $3,095,000 in the third fiscal quarter of Fiscal Year 2013 ("3QFY13"), setting a new record for revenues for any quarter in our history. Software sales increased $692,000 during the quarter. This increase included approximately $300,000 of sales to a major customer that decided in 2QFY14 to move the timing of their global renewal order to our third quarter to synchronize purchases for their internal budget timelines. We expect this long-time customer will now continue to renew in the third quarter going forward. Analytical study revenues decreased by $66,000 during the quarter, mainly due to 3QFY13 collaboration revenues that were not replaced in 3QFY14. Within analytical study revenues, we distinguish between consulting services and collaborations; it was the completion of the funded collaboration to develop the dermal dosing capability within GastroPlus that results in reduced analytical study revenues. Training revenues increase by $20,000.

. . .

  Add SLP to Portfolio     Set Alert         Email to a Friend  
Get SEC Filings for Another Symbol: Symbol Lookup
Quotes & Info for SLP - All Recent SEC Filings
Copyright © 2014 Yahoo! Inc. All rights reserved. Privacy Policy - Terms of Service
SEC Filing data and information provided by EDGAR Online, Inc. (1-800-416-6651). All information provided "as is" for informational purposes only, not intended for trading purposes or advice. Neither Yahoo! nor any of independent providers is liable for any informational errors, incompleteness, or delays, or for any actions taken in reliance on information contained herein. By accessing the Yahoo! site, you agree not to redistribute the information found therein.