In one sentence:
The ChemAlive interface is a cloud-based computational tool that delivers highly accurate ab initio analytics for the chemical industry and academia through its in-house automation algorithm and large-scale computational database.
Our beta release can be played with here:
Please use firefox to run it.
Since our company was founded a few months ago, we have been building our platform from the ground up with the strong belief and drive that ab initio computational chemistry has evolved to the point where its accuracy can be offered to the entire chemical community as a standard tool rather than a skill resigned to use by experts. We have designed our interface, thus, to take rudimentary chemical information provided by its users in the chemical sciences and transform this information into highly accurate metrics for the analysis of molecular properties and chemical reactions through the use of state-of-the-art quantum chemistry tools deployed on the cloud. Thus, the interface enables barrierlesss access to computational chemistry managed through one-click automation. Underneath this utility we aim to build the largest computational chemistry database in the world (using our unique algorithms) that will serve as a storehouse for optimized structures frequencies and wavefunctions at a variety of levels. The goal is to connect this database across the entire community (chemical suppliers, industry, education and academia) to reduce calculation redundancy and to speed up scientific investigation by pre-compution. ‘A’ is already computed and you need ‘B’; half the job is already done for you. That is the idea.
In analyzing the chemical industry, in particular, we have noticed a significant underuse of ab initio computational chemistry due to the success of molecular mechanics and cheminformatics in providing quick and cheap data for large numbers of molecules. We hope to address this directly with our platform by closing the time and effort gap in accessing better quality ab initio data without the need for in-house HPC systems, personnel and data management/admin.
For academia, ab initio computation is becoming standard fare and we hope to increase this trend by providing faster, more efficient, and automated procedures. We will also be automating benchmarking efforts in energy, spectroscopy, acid-base chemistry and electrochemistry or allowing the statistical combination of intermolecular interactions based on a molecular library. Thus, some of our modules will deal with advanced topics in computational chemistry, by providing a more efficient means to treat them. All this is possible with automation.
Finally, in education we will provide access to chemical metrics for students of chemistry so they can engage in more hands-on learning of organic chemistry and, perhaps, even have a hand at designing new molecules and reactions. We will do this through linking our software to textbook curricula. This will all be done with the Schrödinger equation in the background leaving the phenomenological outcomes to the student.
There has been a confluence of factors that have made our vision a real possibility at this present moment in the history of computational chemistry. New methods and code now deliver greater and greater accuracy at less cost in CPU time, especially with the advent of DFT. In particular, many of us have argued that computational methods can achieve greater accuracy than experiment itself. With this accuracy comes reproducible and self-consistent information about fundamental problems in reactivity, interaction and mechanism with direct utility in gaining control of chemical processes. Secondly, the emergence of cloud computing now allows scalable computational resources fit-to-purpose for computational chemistry without the need for housing, administering and maintaining local computing clusters. Thirdly, big data and storage has made the concept of building and serving computational data on millions of molecules a realizable goal. Our aim is to be The Interface.
We are trying to raise 10,000 CHF to build our database on the QM13 dataset including 134,000 organic molecules (CHNOF) to form a basis for interacting with chemical suppliers in-line with our business model. The 10,000 CHF will compliment 60,000 USD is free time we are applying for on the Microsoft Azure system.
If you like the idea, please contribute! Contribution of 100 USD (CHF) or more will be returned through access rights for equivalent time using our interface when it is officially launched and fully operational. Thus, you will get your contribution back in computing services.
In summary, ChemAlive seeks to empower any chemist to perform professional level quantum chemistry without the need for theoretical training and expertise. Ultimately, using its platform architecture and automation algorithms. We will work over five to ten years to develop this interface towards the full automation and prediction of the outcome of any chemical reactions in the forward or retro-synthetic direction as we aim to become a standard primary tool for all synthetic chemists in all chemical industries, academic research laboratories, university classrooms and consultancies. Our interface offers the following unique aspects:
• We focus on systems, not molecules. Our interface is designed to directly deal with questions of reactivity (kinetics and thermodynamics), supramolecular interactions, benchmarking (energies, spectroscopy and electrochemistry) and microstructures (conformations and acid-base equilibria).
• These questions are asked starting from basic 2D chemical formula and the rest is automated. The user only need know what question they want an answer for and how to draw a molecule. Our automation algorithm handles 2D-3D conversion, conformational analysis and verification, optimization, frequency and thermal analysis and, most importantly, method selections.
• We use only open source code to handle the chemistry.
• Our interface connects to an SQL database that is being built through our automation algorithm to contain the entire library of compounds found in most chemical supplier’s catalogs. The database is grown also through user submission based on user agreement (leasing).
Your help at this stage is essential. Thank you for your time.
Peter Jarowski and Laetitia Bomble
Risks and challenges
ChemAlive faces many challenges fit to our grand goals. We are working towards revolutionizing our domain and addressing, eventually, the complete theoretical prediction of the outcome of chemical reactions. Our team has, together, years of experience in computational chemistry and programing working with the top people in this field. Thus, we have a perspective and concept that we are completely confident in. However, there are certain issues that need to come into alignment before our interface can be properly built, deployed and integrated into standard use by chemists in all sectors. We detail these for each sector and discuss how we will address them and give some details of our business model below as it pertains to the database:
In academia, we believe our database would be widely used by expert computational chemists as well as novices. However, many academic labs already have access to high-end computing and software as well as expertise. Our interface must then address something new. In our experience, academics ask lots of question in a given day but often do not have the time to compute everything. A large database of precomputed compounds would allow them to get answers quickly for starter questions with only 2D structural input needed to begin. Additionally, they would be able to contribute to the database through extra computation and build a profile with us through social media type elements of our interface and website. Thus, the database serves as the core of a developing community aimed at sharing ideas and data for the greater good. This commitment to the community would come with reduced subscription costs and increased access rights to the database. The utility in our interface will redirect monies currently paid for standalone software by many academic groups to our more versatile and automated system.
In education, the interface would be completely accessible to novice students and the database entirely pre-computed following the curricula of popular textbooks. In this way we will be able to charge flat fees for student access and maintain a low cost. We would of course reserves the option of including computation rights promoted through advanced problem sets in the curriculum. We would use our extensive academic network to develop new web 2.0 curricula through, for example, the CDX platform.
In industry, we see the most possibilities for revenue but also the greatest challenges. The first challenge is security. Many companies will not want their high-value chemical targets on our systems. We will thus, eventually, be contracting web security companies similar to those used by major banks to secure monetary transactions. For the moment though, industrial clients will be able to keep their calculations out of our database by paying the full cost for database access and interface use. The pricing will be based on CPU usage with administration charges. The second challenge is in convincing the industrial community that ab intio results are more reliable and that our interface delivers them efficiently and with ease at minimum use of time. This will be developed primarily through building our database fit-to-purpose for the R&D divisions of major companies through collaboration and identifying key examples. In particular, a pre-existent database will allow entry into conversations with chemical suppliers who would like to provide more information to their clients about their offered molecules. Sigma-aldrich, for example, has only 40,000 molecules available for purchase; this is a number we can achieve in our database.Learn about accountability on Kickstarter
- (60 days)