Development of a HLA-B*58:01 Allele Screening System for Allopurinol-Induced Severe Cutaneous Adverse Reactions Detection

The HLA-B*58:01 allele has been identified as a genetic marker in the fatal allopurinolinduced severe cutaneous adverse reactions (SCARs) such as Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN). However, the screening of the HLA-B*58:01 allele prior to allopurinol prescription has not been implemented in hospitals, mostly due to the absence of an inexpensive and rapid screening method. In this work, we have developed a HLA-B*58:01 allele screening system which is able to give results in one and a half hour and is cheaper than any other screening instruments available. The system uses the High-Resolution Melt (HRM) method and generates difference melt curves proportional to a specific DNA’s sequence. A positive screening result will be shown if the HLA-B*58:01 allele is detected in unknown samples screened. Twelve HRM experiments were performed to optimize the system and six final optimized experiments are shown in this paper, along with their difference melt curves. The system was successfully benchmarked against the Illumina Eco Real Time machine and results were also validated by Sanger sequencing. The system was also compared against other HLA screening methods available on the market. We propose that the implementation of this HLA-B*58:01 screening system in hospitals will help to decrease the number of SCARs cases and eventually eliminate them in the long run. Patients will thus be prescribed allopurinol only if they test negative for the HLA-B*58:01 gene and those who test positive will be given alternative drugs, ultimately saving lives, time of treatment and overall medical cost. INDEX TERMS Polymerase Chain Reaction, High Resolution Melting, HLA-B*58:01 allele, Severe Cutaneous Adverse Reactions (SCARs), Difference Melt Curve


I. INTRODUCTION
Allopurinol is the first-line urate-decreasing therapy commonly used in gout patients due to its rapid lowering effect on serum urate levels and affordable cost over other drugs. However, allopurinol use may lead to occurrences of adverse drug reactions (ADRs) in patients which subsequently lead to mortality. These ADRs can manifest as mild maculopapular eruptions (MPE), the lifethreatening severe cutaneous adverse reactions (SCARs) which includes Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN), drug reaction with eosinophilia and systemic symptoms (DRESS), as well as AHS [1].
The HLA-B*58:01 allele, which is part of Major Histocompatibility Complex class I molecule, is the most well-established genetic link to allopurinol-induced hypersensitivity reactions [2]. Patients risk SCARs development 80−97 times more in the presence of the HLA-B*58:01 allele [3]. The genetic association between HLA-B*58:01 and AHS has been validated in numerous Asian countries, including Taiwan, Thailand, China, Hong Kong, Japan, Korea and Europe [2], [4]− [15]. Malaysia was shown to have a high frequency of 10.4 % for the HLA-B*58:01 allele in SCARs recorded [4]. Therefore, it was proposed that HLA-B*58:01 screening prior to allopurinol prescription may effectively reduce incidences of allopurinol-induced ADRs and minimize morbidity and mortality [2]. The HLA-B*58:01 screening is likely to influence the choice of urate-lowering therapies in gout management, which will impact not only ADRs rates but also long-term outcomes and treatment costs of this chronic disease. However, the HLA-B*58:01 screening is time-consuming and currently not available in routine clinical practice. In Malaysia, the laboratory turn-around time (LTAT) of HLA Typing (Class I and II) for disease association test in the Institute for Medical Research (IMR) is 10 working days [16]. The LTAT is calculated from a day after a sample received until the report is sent out to hospital (in working days). According to the Laboratory Service Guide from Department of Diagnostic Laboratory Services Universiti Kebangsaan Malaysia Medical Centre, the LTAT is equal to 15 working days [17]. Meanwhile, the LTAT of Pathology Service in the Hospital Ampang is 30 working days [18]. The guidebook of Immunology Lab Service from the Hospital Universiti Sains Malaysia states that the LTAT for the test is 120 days [19]. This causes a serious problem for patients who require immediate treatment.
The current HLA-B*58:01 screening methods applied in HLA laboratories are serological typing, sequence-specific primer polymerase reaction, sequence specific oligonucleotide probes and direct DNA sequencing [20].
The tests require specific and expensive instruments, which are mainly used in medical research rather than in clinical practice. For example, the cost of CFX96 Touch Real-Time PCR Detection System manufactured by BIO-RAD was MYR 179,800.00 (quote at May 08, 2018). Moreover, experienced staff and laboratory specialists are needed to operate the system and run the tests. None of the current methods fulfill the criteria of a fast and easily accessible screening method which can be used in hospitals.
The aim of this paper is to introduce a small, user-friendly and rapid HLA-B*58:01 screening system which can be placed in hospitals to accompany drug prescription by doctors. Only simple pipetting knowledge is needed to prepare the samples and this can be done by trained nurses or doctors and final results obtained within a few hours. Drug prescription will thus not be delayed and patients will be able to efficiently get the right drug for their genetic makeup. The screening results of the developed system were compared to and validated against the existing Illumina Eco Real-time PCR system using the same screening method. The design and development of the screening system has described and the results of testing have been reported in this paper.

A. HRM REACTION MIXTURE
Saliva samples were obtained from volunteers at the University of Nottingham Malaysia Campus (UNMC) and their DNA extracted by using the Oragene (OG−500) DNA kit (DNA Genotek, Canada). The ethics have been approved by the UNMC Ethical Committee. The positive control was obtained from the UKM Medical Molecular Biology Institute (UMBI). A standard 20 µL HRM reaction contained 1× SensiFAST HRM mix (Bioline, USA), 0.4 µM of both primers which target an area of 150bp of the HLA-B*58:01 allele, variable volume of RNAase-free water (Qiagen, Germany) and 500 ng of template DNA, as shown in Table I. A master mix was prepared for three reactions for each DNA sample to be screened, i.e. duplicates for one DNA sample with its non-template control (NTC). The three reactions for the unknown sample were screened alongside similar three reactions for the positive control. The NTC was used to monitor contamination and false positive results in the screenings. This master mix was vortexed, centrifuged briefly and pipetted into the 0.1 ml PCR tubes, followed by the final DNA template addition. The screening system mimics the principles of the polymerase chain reaction and HRM methods to detect the HLA-B*58:01 allele. The protocol defined the different temperature stages and dwell times for the HRM reaction process. Table II shows the HLA-B*58:01 screening protocol. The total time required for the screening was 1 hour 28 minutes 5 seconds.

A. SYSTEM OVERVIEW
The screening system used the principles of the PCR method and HRM method to determine the presence of HLA-B*58:01 allele in the unknown DNA sample. The system was first developed by adopting a standard PCR protocol and then upgraded to the HRM protocol. Fig. 2. shows the block diagram of the screening system. The system consists of the thermal cycler chamber, fluorescence detection chamber, multifunction data acquisition device, personal computer and power management system.
The thermal cycler chamber controls DNA amplification by using the basics of the PCR method. The chamber raised and lowered the temperature of the samples for the different PCR stages, that is, denaturation, annealing and extension. The function of the fluorescence detection chamber was to record the fluorescence signal during the HRM process. The chamber excited and detected the emission fluorescence signal from the samples.
The multifunction data acquisition device controlled the thermal cycler chamber and fluorescence detection chamber. It acquired the fluorescence data and sent it to the personal computer. A customized graphical user interface was developed for the screening analysis and it also displayed the screening result. The power management chamber was developed to convert the alternating current (AC) power supply into direct current (DC) power supply. It supplied power to the thermal cycler chamber and fluorescence detection chamber.

B. THERMAL CYCLER CHAMBER
The thermal cycler chamber was divided into four main components, which were the PCR block, thermoelectric module, NTC thermistor and temperature controller (Fig. 2). The PCR block was designed to hold the six PCR tubes, with a maximum volume of 20 µL, for the HLA-B*58:01 screening process. It was manufactured in a small size and used an aluminium alloy (6082) to improve the temperature ramping rate of the screening system. High temperature ramping rate helps in reducing the time required to complete the screening. Fig. 3. shows the six-well PCR block design. Six small hollows, with diameters of 1 mm, were drilled beside each depression holding the PCR tubes. It allowed the excitation light to pass through the block to excite the DNA samples in the PCR tubes. Two hollows, with diameters equal to 2.17 mm were drilled in the centre of the block. A bead head negative temperature coefficient (NTC) thermistor (Wavelength Electronic, TCS651), with resistance equal to 100 kΩ ± 1%, was placed into the hollow to measure the block temperature. The temperature of the block was directly proportional to the thermal energy stored in the block. A thermoelectric module (Adaptive Thermal Management, APHC-12708-S), with maximum cooling power equal to 71 watts, was used to modulate the temperature of the block. It functions as a small heat pump, moving the thermal energy from or to the block. The excess thermal energy from the thermoelectric module was removed by a water-cooling system (Cooler Master, Seidon 120 V plus). The operation of the thermoelectric module depended on the polarity and magnitude of the current supply to the thermoelectric module. A Proportional-Integral (PI) temperature controller (Wavelength Electronic, PTC10K-CH), connected to the thermoelectric module and was used to control the direction and magnitude of current supply. The PI temperature controller was proposed due to its simplicity and noise immunity compared to Proportional-Integral-Derivative (PID) temperature controller. The derivative-term of temperature control will cause a fast change in the block's temperature due to noise response, which could in turn destabilize the block temperature [21]. Fig. 4. shows the block diagram of the PI temperature controller feedback control system. The PI temperature controller connected to the 16-bit Digital to Analog converter (DAC) of the multifunction data acquisition device (National Instrument, USB-6003). It drove the thermoelectric module to maintain the PCR block temperature, ( ) at the temperature setpoint, ( ) based on the feedback voltage, ( ) from the NTC thermistor. The feedback voltage, ( ) was determined by (1), where ( ) was the thermistor resistance at time, and was the bias current supplied from the temperature controller. The was determined according to the TCS651's datasheet. Based on the datasheet, the temperature of the PCR and HRM protocol fell into the thermistor temperature range of +41 °C to +114 °C, where the was equal to 100 µA. The DAC converted the temperature setpoint, ( ) into an Analog output voltage, ( ) by using (1) and (2). In this case, the ( ) in (1) was replaced by ( ) . Equation (2) shows the Steinhart-Hart equation, where A, B, and C are the Steinhart-Hart coefficients, ( ) is the thermistor resistance and ( ) is the temperature setpoint. The Steinhart-Hart coefficients were obtained from the TCS651's datasheet. The thermoelectric module connects to the output of PI temperature controller. It pumped the thermal energy, ( ) to or from the PCR block based on the direction and the magnitude of the current, ( ) supplied from the controller. The direction and magnitude of current was determined according to the voltage difference between the ( ) and ( ) in the PI temperature controller.
shows the side view of the PCR heating chamber. The thermoelectric module was placed between the PCR block and the water-cooling system. The PCR block is located at the center of the heating chamber. It is held by the Teflon plastic casing. The Teflon plastic has a high melting temperature of 327 °C. It can withstand the PCR block's high temperature during the PCR and HRM process. It was also used as a heat insulator to reduce the thermal loss from the PCR block. A copper plate was placed on top of the Teflon plastic to conceal the remaining exposed area of the PCR tubes. An aluminum alloy heating lid was placed on top of PCR tubes. The heating lid closed tightly with the copper plate to prevent heat loss to the surrounding. Two thermal electric modules were used to heat up the copper plate and heating lid to 120 °C. The high temperature of the copper and heating lid prevented condensation in the PCR tubes. The unwanted condensation could alter the concentration of the VOLUME XX, 2017 1 HRM solution and lead to incomplete DNA amplification and erroneous results.

C. FLUORESCENCE DETECTION CHAMBER
The fluorescence detection chamber consisted of the lightemitting diode (LED), LED driver circuit, silicon photomultiplier and transimpedance amplifier (Fig. 6.). The fluorescence detection chamber was designed without using a mechanical system. The selection of the LED and the silicon photomultiplier depended on the excitation and emission spectrum of the fluorescence dye, EvaGreen TM . The excitation and emission spectrum for the EvaGreen TM dye are 470 nm and 530 nm, respectively [22]. Six high brightness LEDs with glass lens (Wavelength Electronic, LED470L) were used as the excitation light source. The LED470L has a centre wavelength of 470 ± 5 and an optical power equal to170 mW. The LED470L is a compact, energy-efficient light source and can emit light over a wide range of wavelengths compared to the Laser and Xeon lamp. The excitation light generated by LED passes through the excitation filter to excite the HRM mixture. The central wavelength and the bandwidth of the excitation filter was 470 nm and 30 nm, respectively. The LED driver was developed to supply a constant current to the LEDs. The driver was digitally controlled by the USB-6003.
Six low noise, fast, blue-sensitive silicon photomultipliers (SensL, MicroFC-10050-X18) were used to acquire the fluorescence signal from the fluorescence binding dye. The MicroFC-10050-X18 has a peak wavelength and a spectral range equal to 420 nm and 300 nm to 800 nm respectively. Once the HRM mixture had been excited by the excitation light, the EvaGreen TM dye in the HRM solution emitted the fluorescence light. The light was focused by the planoconvex lens and passed through the emission filter to the silicon photomultiplier. The diameter and aperture of the lens were equal to 6.5 mm and 90 % of its diameter, respectively. The excitation and emission filter were used to prevent the overlapping of the excitation light spectrum and fluorescence light spectrum, which could cause a detection error.
Each silicon photomultiplier was connected to an independent signal conditioning circuit. The signal conditioning circuit is a transimpedance amplifier circuit. The transimpedance amplifier circuit was built by using ultra-low input bias current precision amplifier (Texas Instruments, LMP7721MA/NOPB). The resistance gain of the circuit was equal to 200 KΩ. It converted the current signal (fluorescence signal) from the MicroFC-10050-X18 into a voltage signal. The voltage signal was sampled by a 16-bit Analog to Digital Converter (ADC) of the USB-6003.

D. GRAPHICAL USER INTERFACE
A customized GUI was developed in the personal computer for the HLA-B*58:01 allele detection. The GUI analysed the fluorescence data collected from the silicon photomultiplier. The fluorescent data was collected during the HRM process. The acquired data was saved into a text file (.txt). The premelt region and post-melt region of the HRM stage were defined at 83 °C and 90 °C respectively. The difference in background of fluorescence signals were removed by using baseline normalize method [23]. The aligned plot provided a scaled view of data, enabling easy discrimination of sequence variants that display true difference in their melting curve behaviour. The difference melt curve was used to visually determine if the unknown DNA sample has the same melting curve as the positive control (reference sample). The melt curve generated by the difference plot is unique for each DNA sample as it is representative of its DNA sequence. Hence the melt curve of the positive control is used as a reference and compared to the melt curves of other unknown samples. This difference melt curve was plotted by using the normalized fluorescence data of the unknown sample minus the fluorescence signal of the positive control as the y-axis, against the temperature on the x-axis.

A. THE DEVELOPED HLA-B*58:01 SCREENING SYSTEM
The HLA-B*58:01 screening instrument and customized GUI was successfully developed as shown in Fig. 8. Table III shows the specification of the developed instrument. The HLA-B*58:01 screening algorithm was developed by using the difference melt curve method. The pre-melt and post-melt regions of the melting curve were defined at 83 °C and 90 °C respectively. The normalization of HRM signals was performed by using the baseline normalize method. The normalization eliminated the variation between each fluorescence signal due to the variation of internal resistance of the LED. The screening algorithm allowed the user to define which optical channel contained the unknown DNA sample, positive control (reference) sample and the non-template control (NTC). The software allowed the end user to define the positive and negative thresholds. The positive and negative threshold were set at 45 and −45. This threshold was set previously based on literature as well [24]. Fig. 9 shows the two representatives HLA-B*58:01 screening results by using the developed software. The maximum and minimum peak values were determined and displayed in the software. The software displayed the final screening result by comparing the peak values and threshold values. Fig. 9 (a). shows the positive HLA-B*58:01 screening result, where the maximum peak value was higher than the positive threshold, resulting in a negative screening result. Fig. 9 (b). shows the negative HLA-B*58:01 screening result where the melt curve did not exceed any of the threshold values, resulting in a positive screening result.

B. HLA-B*58:01 SCREENING SYSTEM EVALUATION
The system was optimized systematically with twelve screening runs (see Appendix A) performed and benchmarked against the Illumina Eco Real-time PCR system. In this section we show the six final, fully optimized results performed on the developed HLA-B*58:01 screening system and similarly on the Illumina Eco Real-time PCR system. The six experiments consisted of three positive screening results and three negative screening results. Due to the lack of positive DNA samples containing the HLA-B*58:01 allele, we used different colonies of Escherichia Coli (E. coli) bacteria containing the positive control which was previously cloned in for unlimited supply. These different bacterial colonies have a slight difference in bacterial DNA sequence but they all contain the same part of the positive control's DNA. This was verified by sending the plasmids for Sanger sequencing which validated this claim (see Appendix B). Using different colonies of the same positive control as reference during the HRM runs generated different melt curves shapes. As the melt curve generated is proportional to a sample's DNA sequence, this claim is further confirmed. All the six samples used for this screening were sent for Sanger sequencing (see Appendix B) to confirm the presence or absence of the HLA-B*58:01 allele and to validate the developed system's results. We used the different colonies, cultured under different conditions as 'mock positive' results to set a maximum and minimum level of fluorescence as threshold. Similarly, the negative screening results were also computed to define this threshold [27]. A positive screening result is defined as the unknown sample's melt curve being below the fluorescence threshold level and the opposite for a negative screening result. Fluorescence thresholds of around 45 and -45 were defined after the optimization and testing. Active melt regions were fixed at 83 °C to 90 °C for both systems for result generation and analysis. Fig. 10 (a). and Fig. 10 (b). show the compiled difference melt curves from both systems, for the positive and negative screening results respectively. All the melt curves for colonies 1-3 on the Illumina machine have a small difference in the curves' shape and all reach a peak at around a fluorescence level of 70 to 75 as shown in Fig. 10 (a). The peaks also all have the same peak melting temperature of around 87 °C to 87.5 °C. However, the prototype's results in Fig. 10 (a)., for C1 to C3, all adhere to this threshold to show definite positive screening results, proving the prototype's higher sensitivity. The peaks all have a fluorescence level below 45 and the same melting temperatures fitting in the range of 87 °C to 87.5 °C. Similar melting temperatures of the peaks indicate the similarity of samples owing to their specific DNA sequence [25], thus proving the prototype's reproducibility compared to the highperformance Illumina machine. The prototype's higher sensitivity can be attributed to the different light source and detection method used for fluorescence detection and capture [26]. The sensitivity of the prototype is also shown in the conversion of the small fluorescence values recorded in real time to a fluorescence value (×100) for all six samples. This is because the MicroFC-10050-X18 is a single-photon sensitive low-light detector which can detect very low fluorescence signals. Moreover, the prototype was designed and optimized specifically for one screening method only, compared to the Illumina machine, which has a broad range of use. Fig. 10 (b). shows the negative screening results for DNA samples 1-3, where the Illumina machine's results show similar melt curve peak shapes for the negative screening VOLUME XX, 2017 9 results when compared to the positive screening results in Fig. 10 (a). DNA1-3 all have fluorescence levels of 50°C to 65°C and melting points around 87°C to 87.5°C. Thus, a clear distinction cannot be made for the Illumina machine for the positive or negative screening results. However, the prototype's results in Fig. 10 (b). clearly show the three DNA samples' melt curves exceeding the thresholds of 45 and −45. The melt curves generated by the prototype for DNA1−3 all have different melt curve shapes, peaks and melting temperatures as shown in Fig. 10 (b). This goes hand in hand with the basic notion that DNA1−3 are three different DNA samples taken from three different individuals. The latter is not portrayed in the Illumina machine's results. Therefore, the prototype shows more specific and sensitive results in the generation of difference melt curves for positive and negative samples. The prototype can also easily identify the presence of the HLA-B*58:01 allele as a positive screening result, even when using bacterial colonies with only an infinitesimal difference in DNA sequence. This fulfills the criteria of an effective and sensitive screening system.

C. COMPARISON TO EXISTING INSTRUMENTS
The developed HLA-B*58:01 screening system was compared against Illumina Eco Real-time PCR, Rotorgene Q 5 Plex and CFX96 Real-time PCR as shown in Table IV [27]− [29]. The available real-time PCR systems in the market were developed for laboratory purposes compared to the application specific prototype. The real-time PCR systems in the market can perform allelic discrimination (SNP genotyping) compared to the application specific prototype which it developed for HLA-B*58:01 allele screening. Therefore, the block format of the prototype (6well) is smaller than all the other available machines. The prototype used cheaper, commercially available consumables such as 0.1 mL PCR tubes compared to the Illumina Eco Real-time PCR's custom plates and optical adhesive seals, the Rotor-Disc of the Rotorgene Q 5 Plex and the Bio-Rad's PCR Plates of the CFX96 Real-time PCR. These systems used system specific consumables which increased the final screening cost. The well-to-well uniformity of the prototype (± 0.1 °C) was equal to the Illumina Eco Real-Time PCR (± 0.1 °C), higher than the Rotorgene Q 5 Plex (± 0.01 °C) and lower than CFX96 Real-time PCR (± 0.4 °C). However, the average thermal ramp rate of the prototype (4.35 °C/s) was lower compared to Illumina Eco Real-time PCR (5.5 °C/s), Rotorgene Q 5 Plex (17.5 °C/s) but higher than CFX95 Realtime PCR (3.3 °C/s). The dimension of the prototype (220 mm × 200 mm × 360 mm) was the smallest and its weight of 3.6 kg was the lightest compared to all other machines. The small and lightweight prototype is ideal for use as a portable screening system in hospital settings. Furthermore, the prototype's instrumentation cost of MYR 22,285.00 was the cheapest compared to all the other machines. Low cost realtime PCR instruments can overcome the budget constraints and encourage its implementation in hospitals and reduce the allele screening cost.

D. COMPARISON TO OTHER COST OF HLA-B*58:01 TEST METHODS
In Malaysia, the Institute Medical Research (IMR) offers the HLA-B*58:01 test. IMR uses the sequence-specific oligonucleotide (SSO) technique and the cost per sample is MYR 500. In addition, there are several HLA genotyping kits available in the market. However, the cost for HLAgenotyping using these kits is quite expensive although the LTAT is less than a day. At the UKM Medical Molecular Biology Institute (UMBI), the cost for HRM-PCR using Sensimix HRM EvaGreen TM is about MYR 160.00 and the LTAT is around 5 hours. This prototype is an open system, thus the users could choose any kind of QPCR reagents that gives high specificity and sensitivity with reasonable price.

E. HLA-B*58:01 SCREENING SYSTEM LIMITATIONS AND FUTURE WORK
In this work, the prototype was developed for clinical screening purposes and used the HLA-B*58:01 allele primarily in this first stage of development. Although the real-time PCR machines in the market can perform DNA quantification, allelic discrimination and identifying unknown mutations, it is primarily designed and caters for a wide range of research purposes. Our screening system focuses on specific HLA-B*58:01 allele screening in this first stage of development, rather than a wide range of research purposes. However, by changing only the primers and annealing temperature, this system can be used for the screening of any other HLA alleles in hospitals, along with future possibilities of targeting any other genes. A primer pair targeting a length of 150bp and a fixed, optimized DNA concentration were used in this study, as the main focus was on building a specific screening system. Future work will include testing and optimization of numerous length of target genes, along with a range of different DNA concentrations to test the precision, limits and specificity of the system. The foundation of this screening system lies on the generation of HRM melt curves with factors such as primers used, target amplified and annealing temperature easily modifiable, similarly to any other real-time PCR machines. Hence, this screening system should be able to target any other target genes, different gene lengths and multiple DNA concentrations, with proper optimization and testing.
Another limitation of this system is that it currently uses only one type of fluorescent dye and cannot detect other types of fluorescent dyes. The optical excitation and detector of the system were designed for EvaGreen TM fluorescent dye specification only and modification of the optical detection system will be required in order to detect any other dyes. The developed screening system has been used to test for the presence of the HLA-B*58:01 allele in gout patients with different types of ADRs and SCARs by using a positive control from a patient with Stevens Johnson Syndrome (SJS) (see Appendix C). However, only a small number of experiments were done and a greater sample size needs to be tested and validated to truly prove the system's accuracy and sensitivity.
The total time taken for DNA extraction in the laboratory is around 3 hours and this greatly impacts the total screening time for this screening system. This time taken for DNA extraction can be reduced by half by integrating the whole DNA extraction process in a microfluidic chip, as done by G. Samla et al. [30]. Future work will focus on making this screening system a faster screening machine with integrated DNA extraction from saliva/blood, followed by the HRM method and result generation.

V. CONCLUSION
The HLA-B*58:01 screening system was successfully developed based on the principles of the HRM method. The developed system can determine the presence of HLA-B*58:01 allele by comparing it's HRM melt curve to the positive control's melt curve and the screening results were validated against sanger sequencing results. Moreover, the screening system showed more specific and sensitive results in the generation of difference melt curves for positive and negative samples compared to the Illumina Eco Real-time PCR machine. This screening system is portable and inexpensive as compared to all other HLA screening machines available. User-friendliness is showen by the simple pipetting knowledge needed to prepare the samples, which can be done by trained nurses or doctors. Moreover, results are displayed as positive or negative directly, removing the need for any complicated and time-consuming result analysis. Its properties make it ideal for implementation in hospitals in order to provide on-site HLA screening, followed with the correct drug prescription to patients based on their genetic makeup. This will ultimately lead to a decrease and an eventual elimination of SCARs in hospitals and lead the way for the personalized medicine era.  GAGCTCGGGCCCCCACACGTGTGGTCTAGAGCTAG  CCTAGGCTCGAGAAGCTTGTCGACGAATTCAGATT  CGCAGGTTCTCTCGGTAAGTCTGCGCGGAGGCCTTC  ATGTTCCGTGTCTCCCCGTCCCAATACTCCGGCCCC  TCCTGCTCTATCCATGGCGCCCGGGGCTCCGTCCTC  GGACTCGCGGCGTCGCTGTCGAACCTCACGAACTG  GGTGTCAATCACGAATTCTGGATCCGATACGTAAC  GCGTCTGCAGCATGCGTGGTACCGAGCTTTCCCTAT  AGTGAGTCGTATTAGAGCTTGGCGTAATCATGGTC  ATAGCTGTTTCCTGTGTGAAATTGTTATCCGCTCAC  AATTCCACACAACATACGAGCCGGAAGCATAAAGT  GTAAAGCCTGGGGTGCCTAATGAGTGAGCTAACTC  ACATTAATTGCGTTGCGCTCACTGCCCGCTTTCCAG  TCGGGAAACCTGTCGTGCCAGCTGCATTAATGAAT   CGGCCAACGCGCGGGGAGAGGCGGTTTGCGTATTG  GGCGCTCTTCCGCTTCCTCGCTCACTGACTCGCTGC  GCTCGGTCGTTCGGCTGCGGCGAGCGGTATCAGCT  CACTCAAAGGCGGTAATACGGTTATCCACAGAATC  AGGGGATAACGCAGGAAAGAACATGTGAGCAAAA  GGCCAGCAAAAGGCCAGGAACCGTAAAAAGGCCG  CGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTG  ACGAGCATCACAAAAATCGACGCTCAAGTCAGAGG  TGGCGAAACCCCGACAGGACTATAAAGATACCAGG  CGTTTTCCCCCTGGAAAGCTCCCCTCGTGGCGCTCT  CCTGGTTCCGACCCTGCCCGCTTACCGGAATACCTG  GTCCGCCCTTTCTCCCCTTCGGGAAAGCGTGGGCGC  TTTTCTCATAGCTTCACGCCTGTAAGGTATCCTCAG  GTTCGGGGGTAGGGTCGTTTCGCCTCCAAGGCTGG  GGCTGGTGTGCACGAAACCCCCCCGTTTCAGCCCC  GAACCGGCTGGCGCCCTTTAATCCCGGAAAACTAA  TCGTCCTTGGAGGTCCCAACCCCGGGTAAGAAACC  CAAACTTGATTCGCCCACCTGGGGAGGCCAGCCCC  ACTGGGGTAAACGGGGAATTAAACAAAAAGCCAA  GGGAAATGGTAAGGGCGGGGGGCTTACCAAAAATT  TCTTTGGAAAGGGGGGGGGGGCCCTAAACCCTACC  GGGTTTAACCCCTTAAAAAAAAAAACGGGAATTTT  TGGGGTAATTCTGGGGCCCTTTGGGCTTGAAAAAC  CCAGGTTTAAACCTTTTCGGGAAAAAAAAAGAAAA  TTTGGGGAAAACCCTCTTTGGTATTCCCGGGGCAAA  AAAACAAAAAACCCCCCCCCCCTTGGGGGAAAACG  GGGGGGGGGGTTTTTTTTTTTTTGGTTTTTTGGCAA  CAAAAGCCCCCCCAAAAAAAATAAACCCCCCCCCC  CCAAAAAAAAAAAAAAAAAAAAGGAGGGGGTCCC   Colony 2  >1st_BASE_3028840_P2_3_2_M13F__20  ACGTTAATCATTTTGGTTGACTATAAGATACAGCGG  CCGCGAGCTCGGGCCCCCACACGTGTGGTCTAGAG  CTAGCCTAGGCTCGAGAAGCTTGTCGACGAATTCA  GATTCGCAGGTTCTCTCGGTAAGTCTGCGCGGAGG  CCTTCATGTTCCGTGTCTCCCCGTCCCAATACTCCG  GCCCCTCCTGCTCTATCCATGGCGCCCGGGGCTCCG  TCCTCGGACTCGCGGCGTCGCTGTCGAACCTCACGA  ACTGGGTGTCAATCACGAATTCTGGATCCGATACGT  AACGCGTCTGCAGCATGCGTGGTACCGAGCTTTCCC  TATAGTGAGTCGTATTAGAGCTTGGCGTAATCATGG  TCATAGCTGTTTCCTGTGTGAAATTGTTATCCGCTC  ACAATTCCACACAACATACGAGCCGGAAGCATAAA  GTGTAAAGCCTGGGGTGCCTAATGAGTGAGCTAAC  TCACATTAATTGCGTTGCGCTCACTGCCCGCTTTCC  AGTCGGGAAACCTGTCGTGCCAGCTGCATTAATGA  ATCGGCCAACGCGCGGGGAGAGGCGGTTTGCGTAT  TGGGCGCTCTTCCGCTTCCTCGCTCACTGACTCGCT  GCGCTCGGTCGTTCGGCTGCGGCGAGCGGTATCAG  CTCACTCAAAAGGCGGTAAATACGGTTATCCACAG  AATCAGGGGATAACGCAGGAAAGAAACATGTGAG  CAAAAGGCCAGCAAAAAGGGCCCAGGGAAACCCG  TAAAAAAAGGGCCCGCCGTTTGCTGGGCGTTTTTTC  CATTAGGGGCTCCCGCCCCCCCCCTGGACGAAGCA  ATCACAAAAAAAATCGAACGGCTTCAAAGTCCAAG  AAAGGGTGGGGCCGGAAAAAACCCCCCGGAACCA  GGGGAACCTTAATTAAAAGAAATAACCCCAAGGGG  CCGGTTTTTTTCCCCCCCCCCCTTGGGGAAAAGCCC  TTCCCCCCTTCCGGTGGGCGGGCCTTTCCTTTCCCTT  TGGGTTTTCCCCGGAAACCCCCCTTGGGCCCCCGGC  CTTTTTAACCCCCGGGGGAAAATAACCCCCTTGGGT  TCCCCCGGGCCCCCTTTTTTTTCCTTTCCCCCCCTTT  TTCCCCGGGGGGGGAAAAAAGGCGCGGGGTGGGG  GGCCCCGGGCCCCTTTTTTTTTCCCCCTCCCCCAAA  TTAAAAAGGGCCCCTTTTCCCAAAAACCGGGGCCC  CCTTTGGGGTTAAAAAAGGGGGGGTTAAAATTTTC  CCTTTTTCCCAAAAGGTTTTTTTCCCGGGGGGGGGT  GGGGGTTAAAAAGGGGGGGTTCCCCGGGGGTTTTT  TCCCCCGGCCCTTTTCCCCCCCAAAAAGGGGCCCTT  TGGGGGGGGGCCCCTTGGGGTGGGGGGTGGGGGGC  CACACCCCGGAAAAAAAAACCCCCCCCCCCCCCCC  GGGGGTTTTTCCCCAAAGGCCCCCCCCGGGAAAAA  ACCCGGGGGCCTTTGGGGGGCCGCCCCCCCTTTTTA   AAATTCCCCCGGGGGGGAAAAAAAAACCCAATATT  ATTTCCCGCGGGCCCCCCCTTTTTGGGGGAAGGTTC  CCCCCAAAAACCCCCCCCCGGGGGGGTTTAAAAAG  AAAAAACCCCGGGGAAAACTTTTTTAAAATCCCGG  CCCCCCACCCTTGGGGGGGGGAAAAAAGACAAGG  GCCCCCCACCCTGGGGGGGGGAAAAAAACACGGG  GGAATTTTTAAGCCCCCCAAAAAAGCCCCCGAGGG  GGGAAAATTGGGGTTAGGGGGGCGGGGGGGGGGG  CCCTTCACCCCAGAAAAAATTTTTCTTTTGGGGAAA  AAAGGGGGGTGGGGGGCCCCCTAAAAACTTTTCCC  CGGGGGGTTAAAACCCCCTTTAAAAAAAAAAAAAA  AAACCCAGAAAATTTTTGGGGGAAAATTTTCGGGG  GGCCTCTCTCGGGC   Colony 3  >1st_BASE_3028841_P2_1_2_M13F__20  GGGNGTAATTTTTTGTGAACTATAGAATACAGCGG  CCGCGAGCTCGGGCCCCCACACGTGTGGTCTAGAG  CTAGCCTAGGCTCGAGAAGCTTGTCGACGAATTCA  GATTCGCAGGTTCTCTCGGTAAGTCTGCGCGGAGG  CCTTCATGTTCCGTGTCTCCCCGTCCCAATACTCCG  GCCCCTCCTGCTCTATCCATGGCGCCCGGGGCTCCG  TCCTCGGACTCGCGGCGTCGCTGTCGAACCTCACGA  ACTGGGTGTCAATCACGAATTCTGGATCCGATACGT  AACGCGTCTGCAGCATGCGTGGTACCGAGCTTTCCC  TATAGTGAGTCGTATTAGAGCTTGGCGTAATCATGG  TCATAGCTGTTTCCTGTGTGAAATTGTTATCCGCTC  ACAATTCCACACAACATACGAGCCGGAAGCATAAA  GTGTAAAGCCTGGGGTGCCTAATGAGTGAGCTAAC  TCACATTAATTGCGTTGCGCTCACTGCCCGCTTTCC  AGTCGGGAAACCTGTCGTGCCAGCTGCATTAATGA  ATCGGCCAACGCGCGGGGAGAGGCGGTTTGCGTAT  TGGGCGCTCTTCCGCTTCCTCGCTCACTGACTCGCT  GCGCTCGGTCGTTCGGCTGCGGCGAGCGGTATCAG  CTCACTCAAAGGCGGTAATACGGTTATCCACAGAA  TCAGGGGATAACGCAGGAAAGAACATGTGAGCAA  AAGGCCAGCAAAAGGCCAGGAACCGTAAAAAGGC  CGCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCC  TGACGAGCATCACAAAAATCGACGCTCAAGTCAGA  GGTGGCGAAACCCGACAGGACTATAAAGATACCAG  GCGTTTCCCCCTGGAAGCTCCCTCGTGCGCTCTCCT  GTTCCGACCCTGCCCGCTTACCGGAAACCTGGTCCG  CCCTTTCTCCCTTCGGGAAGCGTGGCGCTTTCTCCA  AAGCTCACGCCGGAAGGAATCTCCAGTTCCGGGGT  AAGGCCGTTCGCCTCCAGGCTGGGGCTGGGTGGCC  GAAACCCCCCGGTTCAGCCCGAACCGCCGGCGCCT  TTATCCGGGAAACAATCGGCCTTGGAGCCCAACCC  CGGGAAAGAAACCGAACTTTATCCCCCACTGGGCA  CCCCCCCCCTGGGTAAAACAGGAATAAACCAAAAC  CGGGGGTTTGGAAGGGGGGGGGGCCCCAAAAAATT Table IV shows the Sanger sequencing analysis results which were inputted in the National Centre for Biotechnology Information (NCBI, USA) to perform an alignment with the section of the HLA-B*58:01 amplified by the primer. Presence of the HLA-B*58:01 allele was only identified in the three bacterial colonies whereas the three DNA samples had three other HLA-B identified. This further confirms the results obtained by the screening system.