Program in BASIC to combine the data from two channels of an integrator, and its use in the calculation of residues per 1000 residues amino-acid analyses

Single-channel integrators have been used with gas-liquid chromatography (g.l.c.) equipment and amino-acid analysers for some years and are adequate for many g.l., applications. During amino-acid analysis with the conventional ninhydrin systemmost amino-acids give a blue product but others, notably proline and hydroxyproline, give a yellow product. These products are measured separately, so for manual calculation a two-pen, or three-pen, recorder has always been provided. Attempts to use a single-channel integrator for amino-acid analysis are inevitably unsuccessful. Measuring just the blue channel at 570 nm leads to gross inaccuracies for the small peaks of proline and hydroxyproline; combining the two colorimeter outputs leads to base-line problems, as does the use of an intermediate wavelength. There are now integrators available that enable the output from two or more channels to be integrated, but it is not always possible to combine the results from two channels. Amino-acid analysis results for proteins have been presented as residues per 1000 residues (or per 100 residues)--this method was first introduced as an aid to comparative studies and for this it is necessary to take data for proline, and hydroxyproline ifpresent, from the 440rim channel and combine it with data for all the other amino-acids from the 570rim channel. This paper presents a program that allows the stored data from two channels of a Trilab 3 integrator to be combined to give a full amino-acid analysis expressed in residues/1000 residues. The operator may elect to omit earlier blocks ofdata to reach an appropriate standard, and may also reassign peaks that were wrongly assigned in the automatic mode. With minimal modification the program could be adapted to combine data from two g.l., runs on a single sample and express results in any convenient form.


Introduction
Single-channel integrators have been used with gas-liquid chromatography (g.l.c.) equipment and amino-acid analysers for some years and are adequate for many g.l., applications. During amino-acid analysis with the conventional ninhydrin systemmost amino-acids give a blue product but others, notably proline and hydroxyproline, give a yellow product. These products are measured separately, so for manual calculation a two-pen, or three-pen, recorder has always been provided.
Attempts to use a single-channel integrator for amino-acid analysis are inevitably unsuccessful. Measuring just the blue channel at 570 nm leads to gross inaccuracies for the small peaks of proline and hydroxyproline; combining the two colorimeter outputs leads to base-line problems, as does the use of an intermediate wavelength.
There are now integrators available that enable the output from two or more channels to be integrated, but it is not always possible to combine the results from two channels. Amino-acid analysis results for proteins have been presented as residues per 1000 residues (or per 100 residues)--this method was first introduced as an aid to comparative studies and for this it is necessary to take data for proline, and hydroxyproline ifpresent, from the 440rim channel and combine it with data for all the other amino-acids from the 570rim channel. This paper presents a program that allows the stored data from two channels of a Trilab 3 integrator to be combined to give a full amino-acid analysis expressed in residues/1000 residues. The operator may elect to omit earlier blocks ofdata to reach an appropriate standard, and may also reassign peaks that were wrongly assigned in the automatic mode.
With minimal modification the program could be adapted to combine data from two g.l., runs on a single sample and express results in any convenient form.

Materials and methods
The outputs from the 570nm and 440nm colorimeters of an amino-acid analyser (the LKB 4400, produced by LKB Ltd, Milton Road, Cambridge, UK) are taken via head amplifiers into two of the four channels of a Trilab 3 integrator (Trivector Systems Ltd, Sunderland Road, Sandy, Bedfordshire, UK). All correspondence should be sent to John J. Harding at the Nuffield Laboratory of Ophthalmology.
The manufacturer's operating system and chromatography programs are used for data collection, immediate print-out and storage on mini-cassettes (Philips Data Systems Ltd).
After data collection the manufacturer's basic compiler (Version 5.9) was loaded, followed by the program listed in figure 1.

The program
The REM statements make the program almost selfexplanatory (figure 1). In line 40 the string 'B' represents the three letter codes for the expected amino-acids in order of elution; individual codes can then be extracted using the substring function (for example, line 120). Line 45 sets up the arrays for the two channels with the first array also serving for the assembly of the print-out. Lines 50 to 90 feed in the amount of each amino-acid in the standard analysis: 5nmol for all amino-acids except hydroxyproline (10nmol). The program moves on to line 200 and the operator is asked to load the DATA tape, and then the tape is opened (subroutine 11000 to 11020).
The operator next has the opportunity to bypass any early runs (lines 231-235 and subroutine 17000 in figure 1; see the print-out in figure 2). The operator is then asked to enter the total number of runs before the standard data is read from the tape (subroutine 12000). Subroutine 11130 is also used at this stage--it converts strings that have been stored as numbers back to strings. The computer checks the channel before printing-out the number of the sample used as a standard. Lines 264 and 266 transfer the data for hydroxyproline and proline from the B channel array into the appropriate positions in the A channel.
Lines 267-275, with subroutine 2000, allow new areas to be fed in for selected peaks where the automatic assignment has been incorrect.
Faulty assignment is usually caused by drifting retention times while the analyser is running overnight. Lines 278 to 330 then provide calculation ofresponse factors for the standard and subroutine 100 allows them to be printed-out if required. From line 340 the program allows calculation of each run using subroutines 14000, 18200 and 11130 to read the data, the header and to change stored numbers to strings (line 14035 converts the retention-time units from seconds to minutes). Both channels are accessed before going to subroutine 17000 which bypasses the data for unnamed peaks. The areas for hydroxyproline and 0142-0453/82/0401 0025 $05"00 t' 1982 Taylor & Francis Ltd proline, peaks and 2 on the B channel, are transferred into the main channel array to be the first and seventh peaks respectively (lines 364 and 365). Next comes the reassignment routine (lines 367 to 376 including subroutine 2000), which allows new areas to be assigned to any of the amino-acids and sets the retention times of reassigned peaks to zero as a reminder. Then the concentrations are calculated (lines 379-400), and converted to residues/1000 residues (lines 430 to 730) before the report is printed (subroutine 16000) the array cleared (subroutine 18000) and the tape closed (line 999).
The print-out A print-out demonstrating the options available is shown in figure 2. After the invitation to load a DATA tape the operator was asked if he wished to skip some blocks. He chose to do so as the first two chromatographic separations were poor. The printout gave the sample numbers for the bypassed data and then after being told the total number of runs proceeded with the calibration. The operator chose to reassign three peaks in the basic region (HYL, HIS and LYS) which had been incorrectly assigned; and chose to have the standard concentrations, response factors and peak numbers printed. Moving on to the analysis of a lens protein hydrolysate again, the operator was invited to reassign peaks and did so choosing to set hydroxy-proline to zero and to reassign areas in the basic region (figure 3), taking the areas from the automatic print-out table of unassigned peaks. The results were then printed as shown in figure 3 giving retention times (zero if peak either absent or reassigned), names, areas, concentrations (nmol) and residues/1000 residues.

Discussion
Changes to line 700 would enable the results to be presented in a different form, for example residues/100 residues and with a few extra lines residues/peptide or residues/g could be calculated. Exchange of the routines used to read data from tape would permit the use of this program with other microcomputer systems programmable in BASIC that are used as integrators. This program should have wide applicability to systems where data from two channels must be merged into a single analysis, whether the two channels represent two features of a single run, or two g.l.c., high-performance liquid chromatography or other chromatographic separations.