Date of Award


Degree Type

Restricted Access Professional Report

Degree Name

Master of Science

Field of Study

Cell Biology and Genetics


Graduate School of Biomedical Sciences

First Advisor

John Planz

Second Advisor

Joseph Warren

Third Advisor

Arthur Eisenberg


The AFLP technique at first seems to be a remarkable new technology that can be applied to the growing area of non-human DNA testing. The ability to identify organisms without prior genetic knowledge would be an asset to a field such as non-human DNA testing since not enough research in the area is being conducted. With any new technique or theory in science, intense scrutiny must be used to examine the applicability of the new technology. In the area of forensic science, the severe consequences of a false result extend far beyond the realm of scientific error. Errors make in forensic casework could result in life changing occurrences for the families of not only the victim, but the defendant as well. From this study it can be seen that AFLP as a technique may not stand up to the high expectations of reliability, and reproducibility required for a technique to be adopted into the field of forensic science. Several problems occurred through this study that may prevent this technology from becoming a widely accepted technique in non-human DNA testing. The initial problems with the technique were associated with reproducible results. The first several attempts were conducted under the same conditions, by the same analyst but yielded results that were no comparable. The RFUs of each experiment were inconsistent, not only between samples examined at different times, but samples examined within the same tiral as well. AFLP as a technique is supposedly insensitive to template concentrations however, it has been previously shown to produce differences in the electropherogram when the template is excessively diluted (26). Vos et al. (1995) determined that high dilutions yielding template DNA concentrations below 1 pg could result in irreproducible fingerprints. In this study 27.5 ng of template DNA was added to each digestion-ligation reaction, yet the resulting quantity of amplified fragments varied. These variations in quantities of amplified product could be due to PCR inefficiencies when comparing samples from different trials, but it does not explain instances where duplicate trials were inconsistent with each other (10, 22). When new ligase was introduced the resulting electropherograms did produce considerably higher RFUs for each peak, but the lack of interpretable peaks observed previously may not have been solely due to inefficient ligase. In an inter-laboratory study, Jones et al. (1997) noted that several laboratories encountered problems in obtaining complete AFLP profiles. For several groups, up to 50% of the bands were missing during the preliminary testing. Though this problem subsided with successive attempts, this approach to achieving successful results may not be feasible in a forensic setting. Often the evidence received from a crime scene may be insufficient to allow for multiple testing. In addition, multiple attempts to obtain results may open up areas for scrutiny and attack by the defense counsel. Repetitive testing may appear to be a biased search for condemning evidence against the questioned party, rather than the production of reliable results. Repetitive testing may also not be possible since laboratory reagents and time involved in the production of these results may not be within the constraints of a crime laboratory. In this study, capillary electrophoresis was used to visualize the fluorescent dyes attached to each fragment however, laboratories could use radioisotopes and polyacrylamide gels instead. This method of visualizing AFLP fingerprints is not only costly, but time consuming as well. Conducting repetitive tests in order to obtain a sample with sufficiently intense bands for analysis may not be feasible. These limitations may therefore restricts the use of the AFLP technique from only being conducted in laboratories with sufficient time and funds to conduct repetitive testing as is needed (10). Despite the potential cost in time and funds, the technique was able to produce AFLP fingerprints that were consistent with each other when the electropherograms were compared. The major source of error resulted from the method used to determine the presence of peaks within the designated categories. Since not all peaks crossed the 50 RFU detection threshold, they were not identified by the Genotyper macros. However, when the actual electropherograms were compared, these peaks were present. It has been suggested that to verify whether each peak is present in the pre-designated categories a scan of the electropherogram should be done and any peaks that were not called by the macro should be manually entered into the binary table or should be reanalyzed (Heather Coyle, personal communications). Although this method could potentially aid in the correct genotyping of each sample, it requires a considerable amount of user intervention. A considerable amount of time is needed to examine each electropherogram for the presence of peaks that are below the 50 RFU threshold. Without a redefined interpretation threshold, the analysis of each electropherogram can be highly subjective. Peaks that are relatively low need to be distinguished from peaks that may be associated with background noise. Therefore, in order to eliminate analyst bias a peak detection threshold must be established. Generally the interpretation threshold is established by a validation study of the analysis technique. In this study the lower threshold was previously established at 50 RFU for the instrument being used, but this threshold was insufficient for the recognition of all peaks present during the AFLP analysis. The question then becomes to what extend the peaks can or should be called in order to correctly identify each organism without errors. The exclusion of some peaks could lead to discrepancies, such as those observed during the blind study, which could result in an initial false match or exclusion. The interlaboratory study by Jones et al. found only one scoring difference associated with the absence of one band out of a total of 172 in the AFLP profiles. This error was later associated with experimental errors that incurred during the AFLP procedure. Discrepancies such as this can lead to an erroneous identification of samples that could have severe consequences in a criminal case. At this time, the utilization of AFLP technique for further testing of other organisms such as Cannabis sativa does not seem feasible. A variety of adjustments in the technique need to be addressed before this technology should be further applied to organisms in forensic casework. In order for AFLP typing to be used for forensic casework, major improvements in the technique need to be made. Consistency in obtaining reliable electropherograms with peaks well above the RFU detection threshold must be resolved in order to allow for accurate sample interpretation. This will not only allow for greater consistency between replicates, but will also help in establishing new databases for organisms that are being tested. As with any type of forensic DNA analysis, a database must be established for each organism being tested. Without a reliable database, accurate identification of crime scene evidence cannot be established. A major improvement that is required for the utilization of AFLP typing is the process by which genotypes are identified. Utilizing the macros to identify control and variable peaks to create the binary table was a quick and easy method, however it was not always able to identify the correct genotype. The overlapping of electropherograms in GeneScan ultimately was the best method for accurate identification of the blind samples, but in a real case scenario it would not be feasible to compare each evidentiary electropherogram with those in a database. Advancements in technology will continually introduce new techniques and procedures that could be applicable to the field of forensic science. As with any new technique, the methods and theories must be validated in order to determine whether they can be used in a criminal case. The field of non-human DNA testing is growing and with the advent of new technology such as AFLP, the possibility for establishing a non-human DNA identification method may be on the horizon.


W 4.8 P295A 2005