OSAC

The reaction kinetics of ozonation can be determined after the Osac is reached, and the resistance of the ozone transfer from gas phase to liquid phase becomes insignificant where the concentrations of ozone are uniform in the liquid, in this case, the ozone consumption rate is determined solely by the rate of chemical reaction in the bulk (Charpentier, 1981). The reaction kinetics of dye ozonation, therefore, was studied under this circumstance, and the rate constants of dye decay were determined at various initial dye concentrations and pHs. From Figure 15, the results of dye ozonation at different pHs have shown that the reaction followed a pseudo first-order reaction. Since the oxidizing ability of ozone comes from either molecular ozone or hydroxyl free radicals, the rate of dye disappearance can be formulated as follows ... 

USING OSAC TO KEEP STRUCTURES IN THEIR PLACE... 

Our next major challenge was to integrate a chemical structure handling component into ABACUS. Our progress in using OSAC (Organic Structures Accessed by Computer) for this purpose is the principal subject of this paper. 

By 1986, when we were ready to buy chemical structure software, there was a number of systems on offer. A year earlier the position might have been very different. The main contenders were MACCS, DARC and OSAC, a program derived from the reaction storage and searching program, ORAC. In addition, HTSS and SABRE were of interest although we had doubts about their state of development at the time. 

We selected OSAC primarily because it came closest to meeting this need. A valued bonus was that it was fully compatible with ORAC (Organic Reactions Accessed by Computer) and that upgrade plans of its originator. Dr A P Johnson of the University of Leeds Industrial Services (ULIS) called for the full integration of the two programs to handle both structure and reaction data (ODAC). We were therefore able to provide ORAC for our chemists and to use this as a training vehicle for later OSAC use. 

We have now had OSAC in-house for about five months. Our first task was to load the FPL Compound file, now containing over 14,000 compounds, by converting the records from WLN s. 

In our WLN file salts were recorded as parent compounds with a WLN suffix to specify the salt modification. As we wanted our OSAC file to show salts in their ionised form, and as we proposed to generate the OSAC structures directly from WLN s, we first wrote a program to convert our records to true salt WLN s where appropriate. We then used DARING to convert the WLN file to connection tables. A batch of ferrocenes which are of no current interest was first set aside for later consideration. 

We plan to organise a systematic review of all our structures, to check them for correctness (WLN errors cannot be excluded) and clarity and to add any available stereochemical information. In the meantime, however, we have a useful and generally accurate file. Once we had the WLN file in the appropriate form it was passed through DARING, added to OSAC and screened inside a week. We added OSAC as an option on the ABACUS menu but at this stage there was, of course, no real integration. 

We had previously organised a number of ORAC training courses, conducted by the experts at Leeds, and these served as OSAC courses for many of our chemists. However, we backed them up with a series of short OSAC courses aimed at explaining the conventions used in our file and one or two traps for the unwary. We have no systematic data but anecdotal evidence suggests that the system has been very well accepted by Fisons chemists. 

This system offered many advantages. It was easy to develop from the ABACUS standpoint since, in principle, all we needed to do was add calls to OSAC to draw structures on the graphics plane and make sure we left a space for them on the text plane - a simple job for 1032 and one already anticipated in many of our procedures. 

ULIS were very helpful and responsive in providing the functions we needed and in helping us to use them. Figure 1 shows schematically the link between ABACUS and ODAC (the generalised form of OSAC/ORAC) and VDU-PAC (a set of device independent driver calls provided with ODAC). 

The ODAC and VDU-PAC procedures are represented by the bottom boxes but they cannot be called directly by ABACUS 1032 procedures because 1032 can only call integer functions and doesnot support FORTRAN S return codes. It was therefore necessary to write for each procedure a 1032 external procedure and a FORTRAN jacket procedure to ensure correct communications. The ABACUS-ODAC Macro Library is not strictly necessary but it saves much duplication at the 1032 level by, for example, combining into one procedure all the calls needed to find an OSAC record, to manipulate the two screens and to display the structure as required. 

At this point, which we hope to reach by the end of 1987 or soon after, our major objectives will be met. However, we still have manyplans for ABACUS if the effort can be spared. We are making the Fine Chemicals Directory available on OSAC and expect to be adding several other substantial databases over the next year. We hope to integrate our analytical sample request system with ABACUS to provide substructure access to analytical and physico-chemical data. We are adding procedures to summarise data and, if needed, to interface with RS/1. And we plan to extend the scope of ABACUS to include data other than routine screening data on compounds. 

We chose OSAC as the chemical component of ABACUS because we judged it the system best capable of meeting our prime need for an efficient chemical database system which could provide a high level of integration fiexibility. 

Despite the relatively short time we have been working with OSAC and the sparse... 

Our first months with a program such as this have necessarily involved us in much learning, although, of course, we do not concern ourselves with how OSAC does what it does. We have every reason to expect the pace of development to accelerate as we continue to apply what we have learned. 

The work with OSAC has been carried out almost single-handedly, at the Fisons end, by Peter Millington and I am very grateful for all his efforts. I am also grateful to Peter Johnson, Tony Cook, Kevin Higgins and Paul Hoyle of ULIS for their enthusiasm, help and hard work and to Neill Clift for helpful advice. 

The substructure search feature is similar to that of MACCS, DARC and OSAC, but has the additional possibility of specifying that the substructure is to be in the product or in the reactant or in both or in one but not in the other. (There is a dangerous pitfall here. A search for 0=C-0-C in the product but not in the reactant would miss the reaction of CH3COC6H4OH going to CH3COC6H4COCH3). 

In this presentation, I wish to introduce the related topic of molecular dissimilarity, and to suggest that it may prove to be of equal, or perhaps even greater, value. This concept is not so well-established as its converse. It has been implemented in the Pfizer in-house systems, and has also been investigated at Upjohn. Of the commercial systems, only ORAC/OSAC currently offers a specific dissimilarity option. 

DARC in-house 15,23) marketed by Telesystemes. OSAC (Organic Structures Accessed by Computer) from the ex-Leeds University team ORAC Ltd, appeared rather later 24),... 

Chapter 5 deals with integration and standards both for a reaction indexing systems (ORAC) and for structure management software (OSAC). 

ORAC (Organic Reactions Accessed by Computer) and OSAC (Organic Structures Accessed by Computer) have developed into mature systems for the handling of reaction data and structural data. 

---