Single cell DNA-amplification on-chip

(1)

Single cell DNA-amplification on-chip

L

ITERATURE

T

HESIS

Tom Hopman // 11038691 Master Forensic Sciences

5 ECTS Wordcount: 9074 July 2020 // November 2020 Supervisor: Mw. dr. B. Bruijns Micronit Assessor: Dhr. prof. dr. A. Kloosterman University of Amsterdam

(2)

1

Abstract

Recent developments in the criminal justice system have shown the powerful impact of DNA profiles for prosecution or acquittal in many court cases. Forensic DNA analysis and profiling are challenging aspects in most crime scene investigations due to complex mixtures and the small amounts of DNA that are retrieved. Scientist use STR or SNP profiling but the quality of the profiles is still depending on the quantity of the DNA that is found on the scene as a minimal input of 100pg is preferred. To overcome these challenges whole genome amplification (WGA) of single cells could be performed to amplify the entire genome before downstream DNA analysis to meet up with this requirement. Conventional WGA in laboratories is still a multiday process and requires several manual handlings. To reduce costs and gain time, microfluidic devices could be used for these purposes. As these devices are also able to perform single cell amplification on micro levels, the contamination risk reduces, and DNA mixtures could be analysed more specific. Due to these benefits this study focusses on isothermal multiple displacement amplification (MDA) on several microfluidic devices and the recommendations for forensic implementation.

(4)

3

1. Introduction

Over the past two decades deoxyribonucleic acid (DNA) has been proven to be a powerful and reliable piece of evidence in forensic case studies. The evidential value that comes with DNA could be decisive in many crime scene investigations and court cases (1). Due to the great added value of DNA evidence in the criminal justice system, forensic scientists all over the world are still analysing and developing new techniques to optimize this kind of evidence. Although the current techniques are valid enough for court, still many obstacles could be overcome to optimize the methods and thus the success rate of DNA evidence.

One main challenge in most of the forensic cases is the low amount (less than 100pg) or complex mixtures of DNA found on the crime scene which influences the success rate of the DNA analysis (2). Tracing and collecting biological samples on a crime scene is already challenging due to invisibility, environmental conditions, and contamination but highly important for further steps in the evaluation of DNA traces. Forensic DNA typing and profiling is currently based on several short-tandem repeats (STRs) on the human DNA. This set of STR markers is used universal and contains between the 15 and 25 loci to obtain a specific human profile. The conventional process of DNA evaluation consists of several steps: sampling, lysis and DNA extraction, PCR amplification of the STRs, detection and finally reporting. But to be able to perform such analysis a minimal amount of DNA is required, 100-200pg, which is sometimes unavailable which results in partial or insufficient profiles (1,3). A recent study on the recovery of STR alleles showed that samples with ≥100pg of DNA input generated STR profiles at a mean 21.9 ± 2.6 loci among 25 STR loci, and for others with <100 pg recovered the profiles at a mean of 13.9 ± 4.4 loci (4).

To overcome this challenge forensic scientists are facing intensive laboratory work at which samples could result in a DNA-profile. The process of this analysis could take days causing the loss of valuable time during a police investigation and are sensitive to contamination. Some other more rapid techniques require typically 8 to 10 hours but should still be improved (5). By speeding up this process and to reduce human handling, suspects or perpetrators could be identified within a reasonable timeframe, less than 6 hours, which prevents causing more harm or fleeing (6). Several methods are nowadays available to enhance DNA samples to improve specificity and speed during this process.

Whole genome amplification (WGA) is one of the methods that is capable of amplification of the entire human genome from a small number of cells and overcome the problems that are faced with low copy number DNA. WGA pre-amplifies DNA to a requisite amount before the conventional PCR-cycles (7). WGA uses random primers that can anneal to the starting template and amplifies large sections of the original DNA all at once (2). Therefore, this method can multiply the DNA from a single cell or degraded DNA to the number of picograms that could be analysed in forensic investigations. There are several different methods within the WGA techniques which originate from medical research but are nowadays implemented and studied for forensic purposes. However, WGA is most studied and successful on non-degraded DNA samples from a single contributor, research has shown the potential in forensic case work (8).

To overcome the current challenges in DNA analysis, scientists are developing more rapid, specific, and sensitive methods and techniques for DNA analysis to improve the chain of custody whereby less human sample handling and transport is preferred. One of these current developments is the use of a lab-on-a-chip (LOC) or also called microfluidic devices, in which required laboratory steps are combined into a device of several square centimetres (9). These LOCs are already studied in small numbers for several types of DNA-involved process e.g. PCR, DNA amplification and storage. The LOC consists of enclosed microfluidic channels with internal volumes which could be decreased below 1 µL (10). Due to the small size of these microfluidic devices, temperature and mixing of the liquids could be easily and more rapidly

(5)

4 carried out. Another benefit of these is the enclosed system which reduces the risk of contamination. Due to the small size, in future perspectives, they could possibly be used directly on crime scenes.

1.1 The aim of this literature review

In this literature study the opportunities for WGA (on chip) before STR-PCR (off chip) are studied and discussed. Several WGA methods and microfluidic devices are compared based on their abilities and characteristics which result in an overview and recommendations for the implementation of these methods for forensic purposes based on the amplification yield and time. Thereby the combination of the WGA methods on microfluidic devices will be discussed to overcome the shortcomings of conventional laboratory analysis and opportunities for future research and case work. The search strategy for this study is outlined in Appendix 1.

2.

Whole Genome Amplification methods

In many medical, biological, and forensic fields the amplification of genetic material from a single cell is an important factor. Nowadays there are many methods and techniques that have the ability to copy DNA from small numbers of cells based on earlier developed PCR methods. WGA is one of the techniques and has been used for many applications where only a small number of cells, or only a single cell, is available for amplification. The WGA can perform extensive amplification to yield the required amount of DNA for downstream applications (11). The power of WGA is the ability to generate useful DNA out of a few picograms from a single human cell.

The difference between the more conventional PCR and WGA methods is the specificity of the amplification and the amount of DNA amplified during the reaction. The PCR can duplicate specific parts of the DNA due to DNA primers that are chemically synthesized for specific parts of the genome (12). After 20 to 30 cycles specific parts of interest are amplified causing them to be analysed for further purposes like STRs. Due to the specificity of the method, PCR is widely used in forensics to obtain enough DNA for STR-profiling and to identify individuals. But as mentioned earlier, to start the PCR of STRs enough DNA should be available which could be achieved with the use of pre-amplification with WGA. WGA uses more universal primers to amplify larger (12-1000kb) and more parts of the DNA than the more specific PCR method (<10kb) (11–13).

Several methods are used worldwide, but this study discusses the most used: primer extension preamplification PCR (PEP-PCR) and degenerate oligonucleotide-primed PCR (DOP-PCR); multiple displacement amplification (MDA); multiple annealing and looping-based amplification cycles (MALBAC) (14,15). Although all these methods are promising for forensic purposes, this literature review will mainly focus on the application of MDA on microfluidic devices as this is the only isothermal method which is beneficial for on-spot forensic case work.

2.1 Primer Extension Preamplification and Degenerate Oligonucleotide-Primed PCR

One of the first attempts to amplify the whole genome was based on the PCR and has been studied for the past 30 years. Due to the specificity of the PCR-primers scientists have tried to design these to cover the whole genome during the PCR process. PEP-PCR was one of the first methods that attempted to amplify a whole genome from a single cell but yielded low genome coverage (15). After these first attempts with PEP-PCR an improved version was developed that used degenerate primers, the DOP-PCR.

The DOP-PCR makes use of thermal cycling for denaturation (94°C), primer annealing(56°C) and further strand extension(72°C) (16). The primers that are used have a combination of a random sequence of six base pairs at the 3’-end and a target sequence at the 5’-end. Due to this variety in sequences the

(6)

DOP-5 PCR covers larger parts of the genome than the PEP-PCR, but still only 39% (15). This research also showed that the DOP-PCR has a higher false positive rate (1.0x10-3 _{) than newer developed techniques like MDA} (1.0x10-4 _{) or MALBAC}₍₂_.0x10-4 _{). These false positives arise as a sequencing or amplification error after} amplification and are compared with the original DNA sequence. The difference could probably be explained by the different primers that are used in these methods.

2.2 Multiple Displacement Amplification

MDA is one of the methods that can perform WGA, it can rapidly amplify small amounts of DNA in reasonable quantities. Over the last ten years MDA has changed the way DNA sequencing could be performed due to several different advantages (10,17). The more conventional, PCR based methods create billions of copies of a short part (under 10kb in length) of the genome, while MDA is able to generate an average of 12kb in length up to over 100kb (12). The difference in the length of MDA amplicons is caused by the random hexameric primers that are used and the use of phi29 DNA polymerase. This enzyme has a high affinity for single DNA strands and is therefore a suitable component for the MDA. The hexameric primers consist of a 6 bases long randomly given sequence that could anneal at several parts of a DNA sequence and initiate synthesis (18). The strand displacement reaction synthesizes new strands from the multiple copies from each template all at the same time (Figure 1). The combination of the primers and the phi29 DNA polymerase makes the MDA a productive and successful method for WGA.

Another difference between MDA and PCR techniques is the absence of thermal cycling. Conventional PCR starts at a temperature of 94-96°C for denaturation, then cools to ~68°C for annealing and ends at approximately 72°C during the elongation. The MDA is an isothermal process which amplifies DNA at a constant temperature of approximately 30°C which makes the reaction less sensitive for technical errors. As during MDA all these cycles should also be run through, the phi29 DNA polymerase is able to debranch the double stranded DNA without any thermal changes.

Figure 1. A schematic overview of the multiple displacement amplification method. The green arrows indicate flow over time and the annealing of more primers and phi29 DNA polymerases (15).

(7)

6 Several research groups have studied the specificity and bias of the MDA compared to the other WGA methods. The bias is based on overrepresentation and underrepresentation of different parts of the genome. Compared with other PCR-based methods the MDA has a far less sequence-dependent amplification bias on most regions of the genome (17). The study of Lasken showed a 70-75% genome coverage with MDA of 10 E. coli loci which indicates the loss of several parts of the genome. Although MDA creates some bias it already is reduced by three to four orders of magnitude compared with DOP-PCR, which results in more reliable and complete genome coverage (13). Due to the exponential and nonlinear amplification process from the template the method is still sensitive for sequence-dependent bias like the exponential PCR methods (15). MDA is commonly used and widely applied but requires a good quality template which reduces the usefulness in forensic applications where mixtures and incomplete samples are common. As a result of dropout, it could lead to homozygous loci instead of heterozygous loci during an STR analysis. Single cell genome amplification could be a solution for this challenge.

2.3 Multiple Annealing and Looping-Based Amplification Cycles

In 2012 a new WGA method was reported where PCR and MDA in a complete quasi-linear preamplification are combined to overcome the bias with regular MDA (19). As PCR and MDA both amplify the template strand and copies of that, MALBAC is able to only copy the original genomic DNA from the start of the amplification process (Figure 2). The primers that are used consist of 8 random nucleotides at the 3’-end and a common 27-nucleotide sequence at the 5’-end which could even hybridize to the templates at 0°C (13). But like the PCR method, MALBAC uses thermal cycles to multiply the DNA

Figure 2. A schematic overview of the multiple annealing and looping-based amplification cycles. The looped full amplicons on the right are the final product of MALBAC (20).

(8)

7 with the use of DNA polymerase with strand displacement (Bst DNA polymerase). To anneal the primers and start amplification, the temperature is elevated to 65°C and semi-amplicons between 0.5 and 1.5kb of length are synthesized. At a temperature of 94°C the semi-amplicons are melted off from the DNA template to denature. Due to the generated complementary ends of the complete amplicons in the next cycle, at 58°C these can form a loop which makes them unattainable for further amplification. Due to this method cross hybridization is reduced and more original DNA will be synthesized which reduces bias (19). After five cycles of 6.5 minutes, the MALBAC can provide nanograms of DNA from a lysed cell which is enough for further purposes like PCR and STR profiling in forensic case work.

This new method of WGA has been tested in several studies to compare the MALBAC to the existing methods based on PCR and MDA separately. The original study was performed on single SW480 cancer cells and achieved an 85-93% genome coverage with MALBAC in contrast to 72% with MDA (19). These results were reported after the analysis of single nucleotide variance in the DNA. The bias was reported in numbers of allele dropout rate which was 65% for MDA and ~1% for MALBAC. A different study on single sperm cells and blood tissue was performed to compare the MDA to MALBAC (13). This study concluded that MALBAC provides less amplification bias with more genomic coverage based on single nucleotide variation, ±90% for MALBAC and ±45% for MDA.

Although these studies conclude promising results, other studies show more disunity in their research. Studies on BJ primary human foreskin fibroblast, a diploid human cell line, with available MALBAC and MDA kits showed similar allele dropout rates and unmappable parts of the genome (15). Also, the MDA kits were able to show less false positives, less allele dropout or sequence errors than MALBAC (2.0x10-4 _{MDA; 4.0x10}-4 _{MALBAC), probably due to the high affinity of the phi29 DNA} polymerase. The usability for single cells was in favour of MALBAC while MDA showed better results for a mixture of a few cells. Although some imperfections, MALBAC showed some great potential for WGA due to the ability to amplify only the original templates to reduce the bias. Data have shown that from single cells reliable micrograms of DNA could be generated for analysis further in the process (20).

2.4 Whole Genome Amplification before PCR

As mentioned previously, in forensic case work and databases human profiles are based on the STR markers on several loci, regions on the genome. There are several commercial kits available that can obtain a profile within three hours with direct PCR, rapid PCR, and microfluidic devices (5). As recent developments have shown that the time at which these devices can obtain STR profiles reduces, the required amount of DNA is still a constraint. These study have shown that the GeneAmp PCR system 9700 can generate a full Y-STR profile within one hour but still needs 125pg of DNA as input (21). Another study showed the ability to generate full STR profiles under one hour with the SGM Plus™ kit with the amount of 1ng of DNA input, which is way more than generally sampled from forensic scenes (22). With these recent developments the demand for a reliable WGA method increases as less than the required DNA is obtained in most of the forensic cases (2).

Methods like MDA, MALBAC and DOP-PCR could amplify small amounts, or a single cell, into the required amount of DNA that is required for STR-PCR. The main advantage to using these methods before STR-PCR is that, in optimal conditions, the biggest part of the genome is amplified without the loss of DNA material. With these methods the PCR-STR starts with the same original material but with a larger amount of DNA, enough for STR profiling. If the WGA is completed successfully there should be enough DNA that could be used for STR-profiling and storage. STR-PCR is an irreversible process where the deposited DNA is used for profiling. If WGA is performed well, enough DNA should be available for both purposes.

(9)

8 The disadvantage of the WGA before PCR-STR is the bias that comes along with the methods. Even though scientists can reduce the bias within the WGA methods, there is still a drop-out rate. This could lead to inconclusive results and wrong interpretations of the STR-profiles. To achieve as much genome coverage as possible with WGA, this bias does reduce. As a single human cell contains 6pg of DNA, the WGA should be able to amplify this small amount. Research has shown that also with WGA the input amount of DNA has effect on the results (2). The MDA kit GenomiPhi™ and PCR-kit GenomePlex™ both showed that with an increase of DNA input (0.01, 0.5 and 1 ng) the bias decreased, and the genome coverage increased.

The effect of MDA on the proportion in DNA mixtures have shown that there is a positive effect of WGA before STR-PCR in unequal mixtures (23). This study evaluated the Illustra™ Single Cell GenomiPhi™ DNA Amplification Kit which was used before the conventional STR-PCR. They concluded that the drop-out rate was improved using the MDA amplification and that complex mixtures were more imbalanced which resulted in more successful DNA profiling.

This development could improve the analytical threshold that are required for forensic STR-profiling. The threshold where determined on low input DNA (60pg), high input DNA (1ng) and the associated background noise form these samples to determine if a profile is valid enough. If single cell WGA could be performed before STR-PCR, background noise could be reduced which results in more reliable STR profiles on low input samples (24).

3. Microfluidic devices for DNA analysis

Recent developments concerning DNA amplification, detection and storage are now studied in several forms and methods. One of these new developments is the microfluidic devices and showed new potential towards the facets of DNA studies. The ability to integrate standard biological processes into microfluidic systems makes them a viable method for DNA analysis in forensic case work. Since these devices can extract DNA from single cells, protocols from outside the device could be used without the complexity of conventional devices.

The conventional process steps are already integrated in microfluidic systems (Figure 3). As in several studies the capabilities of these individual steps are well determined, there are also microfluidic systems that have them all integrated in one device (25). Although there are many microfluidic systems that are able to perform multiple steps of the DNA analysis some challenges have to be overcome yet.

Figure 3. An overview of the process steps for forensic DNA analysis in the top row. The other two rows show the steps of conventional techniques and LOC’s (9).

(10)

9 Several microfluidic devices have shown their capabilities to perform on-chip PCR reactions before STR analysis. These systems have been shown to reduce analysis time, reagent consumption and improve portability due to their small size which are beneficial in forensic case work where time plays an important role.

The device of DuVall et al., can generate STR profiles from 10 different DNA regions within 45 minutes with 15 minute on-chip PCR with the PowerPlex® Fusion Master Mix with an input of 5ng (26). For heating and cooling they used an external heating component from Peltier. Another study of Han et al., showed that with their microfluidic device they were able to obtain a full STR profile on 15 different loci with an input of 3.57ng (27). This device was able to perform all steps of DNA analysis on-chip in about 2 hours and was called the “sample-in-answer-out” method. This device also used external heating with their custom-made microsystem.

Both devices could be used for straightforward PCR and DNA analysis from a single contributor with the required amount of DNA but are unfit for more complex crime scene samples as discussed before. To overcome these challenges isothermal WGA could be performed before PCR and STR-analysis to obtain more compatible DNA for these downstream purposes.

For this literature study the abilities of microfluidic devices are discussed regarding their ability to perform WGA and their potential for forensic case work. WGA on-chip should be performed as a pre-amplification step for off-chip PCR to obtain more compatible DNA for further analysis. Because in most cases sample material is limited, the potential of single-cell amplification will be highlighted.

3.1 Chip materials used for WGA

As in conventional laboratories most WGA reactions are performed in well plates or tubes of polypropylene, most microfluidic devices were mainly made of glass or silicon. Although these materials have great properties for on-chip DNA amplification, recently more advanced microfluidic devices are made from plastic materials such as Polydimethylsiloxane (PDMS) or cyclic olefin copolymer (COC). The materials that are chosen for on-chip WGA show method dependent variation as the materials all have their own characteristics (28). As in forensic case work contamination should be prevented, microfluidic devices should preferably disposable, which also entails production costs.

Glass is a widely used material for on-chip analysis for biological samples. The transparency, high biocompatibility, hydrophilic features and rigid support are characteristics that makes glass an important chip material which is often combined with silicon or plastics (29). Although these features, the production process requires cleanroom environments and specific equipment which makes glass a relatively expensive material for LOC’s (30).

Silicon is another commonly used material in microfluidic devices due to its positive characteristics for DNA amplification. Besides the shared hydrophilic properties with glass are there some significant differences (31). Silicon has better thermal conductivity than glass, so more efficient heat transmission to the internal parts of the chip where it is required. One of the disadvantages of silicon is the non-transparency for substance detection with UV-light or controls with fluorescence (30).

In the past 20 years PDMS have become the most common microfluidic substrate in use in academic labs due to the relatively low costs, fast fabrication and ease of implementation (31). The use of PDMS as material for microfluidic device has some advantages over glass or silicon. As first, it is easy to produce and keeps its flexibility through cooling (-50°C) and heating processes (200°C) (32). PDMS has a high transparent property so fluorescent signals could be detected through the microfluidic chip (28).

As the thermal conductivity has been reported as low, it is not efficient to transfer heat towards the liquids on the chip but keeps the internal temperature more stable due to low energy loss (33,34). As PDMS

(11)

10 is a hydrophobic and permeable material, several molecules and components could be adsorbed which could inhibit amplification (28,35). To overcome this problem, several PDMS devices are precoated with Bovine Serum Albumin (BSA) solution which prevents inhibitory enzymes from sticking to the channel surface where these enzymes can interfere with amplification reactions (36,37).

4. Studies on microfluidic devices

As several studies have performed off-chip WGA with successful results, new devices are arising to improve DNA amplification of the entire genome with the use of microfluidics on-chip. These devices bring several beneficial characteristics to the studies on DNA analysis (38). One of the most important benefits of these devices is the increase in the control whereby the sample is processed. Due to the small amounts of reagents that are needed and the microliter volumes of the chambers and channels, the DNA from a single cell could be amplified in a controlled environment and a single device. A second effect is the decrease in contamination risk, due to the contained channels and chambers of the device, which is an important section for DNA analysis in forensic cases (9).

Several groups have studied the abilities of microfluidic devices to perform DNA analysis in separated or combined steps, with or without WGA as an amplification step. For this review several studies are discussed which have performed full WGA on-chip e.g. cell capture, cell lysis and DNA amplification. Most of them use the previously discussed MDA method due to the ability of isothermal amplification. As the microfluidic devices are small, and therefore easy to heat or cool, several studies have shown the ability to perform methods that are based on thermal cycling like MALBAC, this review discusses one as a comparison to the MDA. While most of the microfluidic devices are still used for medical purposes, the devices could be compared and perhaps implemented for forensic DNA analysis. Most of the devices work on single cell amplification which could be beneficial for forensic cases whereby small amounts or mixtures of DNA are collected from crime scenes. If single cells are isolated from a mixture, the discriminative power between multiple contributors increases as separate DNA profiles could be reached.

4.1

On-chip MDA before off-chip PCR

The study of Li et al., on specific oncogenic mutated cells, used on-chip MDA before off-chip PCR and sequencing at single cell levels (39). They developed a microfluidic chip that was able to separate single cells simultaneously to identify, lyse, and perform MDA amplification. To identify different cell types, the mutated oncogenic cells were pre-labelled with fluorescein antibodies to detect these with a fluorescence

Figure 4. The left part of the picture shows the 3 layers that are used, the cells process from the top to the bottom layer. The first layer traps single cells; the second layer consists of the valves; and the third contains the MDA chambers. On the right, the whole device is shown (39).

(12)

11 microscope. This pre-labelling allowed them to keep track of which

cell type was in which channel of the microfluidic device before the amplification.

This device was separated in 3 different levels which all have their own process, the chip contains 20 separated sub-channels and chambers (1.2nL volume each) in the bottom layer where the MDA reactions take place (Figure 4). The chip was fabricated of PDMS with 100μm wide channels. Valves are used which are manually monitored to open and close the channel. This to control the reagents (or air) that enter or leave the device and chambers which are controlled by 5 separate micro valves (Figure 5).

At first the cell-mixture was pumped into the device with a flowrate of 1 μL/min into the main channel, where cells enter the first part of the processing sub-channels. They also studied a 0.5 μL/min flow rate which also resulted in successful single cell trapping, a flow rate of 3.0 μL/min resulted in multiple cells per chamber which was undesirable. After a single cell was trapped, they lysed it with alkaline reagent for 5 minutes at 65°C before moving on to the second chamber in which the cell lysate was neutralized with a neutralization buffer for 1 minute.

MDA was performed in the third chamber of the sub-channel with the use of 48nL amplification reagent, REPLI-g Single Cell Kit from Qiagen. Due to the valve membrane, active mixing of the

amplification reagent and the neutralized cell lysate could be performed for a homogeneous substance. The last chamber of 48nL was used for the MDA reactions and was carried out at 30°C for 3 hours. After this, from the top of the device, the amplification products were collected and 1μL was used for further PCR and downstream STR analysis.

The Sanger’s method was used to perform further DNA sequencing on specific regions of the genome. The Sanger’s method is a PCR sequencing method that uses dideoxy nucleotides and fluorescence to determine the sequence of a DNA region. This method requires an input of 250-500 pg of DNA. Although this research did no in-between quantification of the DNA, after MDA amplification was performed, they were able to accurately detect mutations of the oncogenic cells. The combination of on-chip MDA and off-chip PCR was proven as a successful combination in the analysis of single cells for STR analysis, whereby MDA amplified the sufficient amount of DNA (39).

4.2 Microfluidics with button valve design

Yang et al. also studied the possibilities of microfluidic devices for WGA before qPCR on tumour cells to improve treatment options as these are associated with specific genes (36). The chip that was developed consists of 8 separate chambers in which each a single cell was captured for MDA amplification (Figure 6). This microfluidic device was manufactured by multilayer soft lithography techniques to form the shape of the PDMS device. Several layers of PDMS were bonded together to create three circular chambers (1.2nL) and one square chamber for Figure 5. An overview of the third layer of the device. The green lines indicate the valves, the arrows the direction of the flow. The white circle is the top of the device where the MDA products are collected (39).

Figure 6. The downstream direction of the microfluidic device with button valves indicated with a red circle. The blue square is the MDA chamber (36).

(13)

12 amplification (20.25nL).The device was finished with a pre-cleaned glass slide and the channels were pre coated with BSA-solution (0.5mg/mL). The flow through of the device is controlled by button valves in which the reagents could mix more efficiently as passive diffusion has not been found efficient after analysis with qPCR. The process started by the isolation of single breast cancer cells into the first chamber due to fluorescence targets on the cells. In the first chamber cell lysis was performed with 1.2nL (0.2% EDTA and Tween-20) for 10 minutes while pressed to the second chamber. After cell lysis, neutralization was performed with 1.2nL lysis buffer (1 M HCl and of 1 M Tris·HCL, pH 7.5) in the second chamber and mixed for 10 minutes due to the button valves.

As a final step the cell lysate was transported to the square cell of 20.25nL and mixed actively with the Illustra GenomiPhi V2 DNA Amplification Kit. The final MDA amplification process took 2 hours at 34°C on an AmpliSpeed Slide Cycler, an off-chip heating device, before the samples were collected form the chamber. The amount of DNA that a single breast cancer cell consist is equal to normal cells (6-7pg). After the MDA amplification of this microfluidic device, about 8.27ng DNA was obtained which indicated a 1000-fold amplification. After on-chip WGA they performed off-chip PCR on a Bio-Rad CFX Real-Time system for 70 min at 30°C with 1uL of the amplicons. As downstream analysis, they performed qPRC to assess the quality of the amplicons where they concluded a 33% template-specific amplification and successful detection for targeted breast cancer genes.

4.3 WGA of single epithelial cells

The study of Li et al. used similar steps in their protocol for DNA amplification with the use of REPLI-g Single Cell Kit from Qiagen with a designed microfluidic device (29). They studied intact epithelial cells from snap-frozen postsurgical human endometrial tissues, a cancer cell line, on a PDMS microfluidic device with 12 separate reaction chambers. Before adding the cells to the chip, they were labelled with green fluorescence proteins as markers. The flow system shows similarities from previous discussed devices. There are separate chambers in which single cell isolation, cell lysis and DNA amplification are performed (Figure 7). Micro valves were used to open and close the chambers in between steps or for mixing the reagents with the cell. Further specifications of the microfluidic device were not mentioned in the literature.

Laser tweezers were used to manually move the labelled cells into the first chamber of the device with a microscopic view to see the GFP-labelled cells. After successful isolation D2 lysis buffer was added to the cell and was incubated at 37°C for 10 minutes. The neutralization was performed after full lysis by a neutralization buffer at room temperature to determinate the DNA. For the MDA amplification the REPLI-g Single Cell Kit was used while the chip was placed on a hotplate at 32°C for 16 hours. As standard protocol for this chip suggests an incubation of 8 hours on 30°C for standard PCR tubes, several studies have shown that the chosen settings of the parameters are successful.

To finish the amplification, the chip was heated to 65°C for 3 minutes and cooled on ice afterwards. Figure 7. The on-chip processes of the device where MDA was performed during the DNA amplification step (29).

(14)

13 Results showed a >80% amplification success rate, percentage of the whole genome, with 30-50ng amplified DNA from a single cell. The total reaction gain of the MDA samples was estimated to be 3000-5400-fold. They estimated the error rate (mistakes in amplification) of the MDA reaction at 1.6x10-5_which is comparable with off-chip MDA reactions.

4.4 MDA with SISSOR concept

The study of Chu et al. took a different approach to study and sequence the haplotypes of single human cells (37). They amplified 3 different cells from the same human fibroblast cell line. To overcome the problem of high false-positive calls per genome and short sequences, that could arise with commonly used DNA polymerase and high-throughput sequencing, they developed a new method. The single-stranded sequencing using microfluidic reactors (SISSOR) was developed for accurate single cell DNA sequencing. A microfluidic device was used to separate the double stranded DNA into randomly partitioned single strands and for further amplification using MDA. The downstream purpose of this study was to identify single nucleotide polymorphisms (SNPs) for haplotyping.

The microfluidic device that was used for this study was fabricated with the use of soft lithography. The top layer includes the domed flow channels of 12μm and 20μm, which was later bonded to a 25 μm-thick bottom layer with valves. Both layers were made of PDMS and were bonded to a glass slide. The 24 separated rectangle MDA reaction chambers have a width of 60-65μm and have a volume of 20nL. After finishing the separate parts of the device, they were treated with oxygen plasma in a UV-ozone for 4 minutes to desorb leftover organic products from the surface which could have been occurred during manufacturing. They also used 0.1 % BSA as a precoating on the surface of the amplification chambers to improve the MDA reaction.

The device was designed to perform each step of the process in separate chambers (Figure 8). The first unit isolated a single cell with the use of micro valves from the cell suspension that was injected on the chip. The lysis was performed in the mixing chamber by adding alkaline lysis solution (400 mM KOH, 10 mM DTT, and 1% Tween20). The peristaltic PDMS pump mixed the cell and the lysis buffer in the ring circle for 10 minutes. The alkaline solution in this step also separates the double stranded DNA to single

Figure 8. A detailed overview of the microfluidic device wit SISSOR concept. The first part figures the cell capture; middle part the cell lysis and mixing of the reagents; the right part shows the 24 separate chambers where the MDA reactions are performed on ssDNA from a single cell (37).

(15)

14 stranded DNA (ssDNA) parts of the genome. The lysed cell was then pushed onto the MDA 24 chambers with the use of air and a rotary pump, the chambers closed with micro valves.

The first part of the MDA chamber neutralizes the ssDNA and reagents in the solution with the neutralization buffer (400mM HCl, 600mM Tris-HCl, and 1% Tween20). Secondly, the fragments were pushed down to the next chambers (~20nL) where amplification by MDA is performed. The MDA amplification mix was self-designed and consists of an MDA buffer, 84μM 3′ phosphorylated random hexamers, 2mM dNTPs, 150μM dUTP, and 0.84μg of Phi29 DNA polymerase. To carry out the amplification, the device was incubated for 15 hours at 30°C on a hot plate. As a final step, the amplicons were collected with a pipet tip and incubated at 65°C for 10 minutes to end the amplification reaction.

The results of the on-chip amplification were compared with the whole genome of the human cell line which has been sequenced successfully several times with off-chip sequencing methods. The three single cells that were processed with this microfluidic device were able to cover 92.6%-98.8% of the entire genome and an average of ~64% per cell. The difference between single cells and the average of those combined could be explained by the loss of several parts of the ssDNA fragments while they were transported to the reaction chambers from the mixing circle. The estimation of ssDNA amplification with MDA was 60,000 times per fragment. The uniqueness of this device lays in the fact that the DNA of a single cell was amplified while divided over 24 separate chambers.

4.5 Micropillar arrays and MDA

A different design of a microfluidic device was studied by Tian et al. on individual human cancer cells from the HeLa cell line, an immortal cell line commonly used in genome studies (40,41). They designed a microfluidic device that performs single cell isolation, purification, and amplification due to the use of micropillars. This should reduce the amplification bias that occurs during WGA with MDA or other methods. The extracted DNA gets immobilized due to these pillars and makes it available for isothermal amplification with MDA. The difference between the previous discussed devices is the amplification during continuous fluid flow instead of the chambered reactions which require complex channel geometries and valves. To avoid this, they developed the genomic amplification via micropillar arrays (GAMA) microfluidic device.

The chip was fabricated using standard photolithography techniques with PDMS as material, the depth of the fluid channels was 20-25μm. The PDMS plate was bonded to a silica wafer and contained a single input port and 10 separate channels in which each a single cell could be captured and amplified. The first part of the channel consists of a single cell capture region followed by the micropillar region with

Figure 9. The flow direction of the cells is indicated with the blue and orange arrows. The cell capturing part is at the start of the micropillar area where the micropillars are placed in a specific composition. As only one cell could be trapped, the others will flow through the side of the device (40).

(16)

15 different compositions (Figure 9). The micropillars on the cell capture region were placed 1.5μm apart from each other to trap a single cell. The flow rate during cell loading must be above 2μl min-1 _{to prevent cells to} adhere to the surface in the microchannels.

After cell trapping, the flow rate was reduced to 0.5μl min-1_{and the microchannels were rinsed with} phosphate-buffered saline for 5 minutes to wash off the uncaptured cells. To lyse the cell, 6M guanidinium thiocyanate in water was flowed for 5 minutes through the device before washing with 100% ethanol for 5 minutes. As a result, the chromosomal DNA is entangled on the pillars and the other components from the cell are washed off e.g. proteins and lipids (Figure 10).

Before MDA was performed, the double stranded DNA was denatured with D1 buffer for 8 minutes and flushed with N1 buffer for 15 afterwards which result in ssDNA. The REPLI-g UltraFast Mini Kit from Qiagen was used for the MDA reactions which were performed for 3.5 hours at 33°C with a flow rate of 0.5μl min-1_{on a hot plate}_{. The amplification products were captured in the output reservoir, collected} with a pipet, and inactivated in a PCR tube for 10 minutes at 65°C before further analysis. The amplification yield was measured using the Qubit 2.0 Fluorometer and dsDNA HS dye kit (1:200) and found close to a 10,000-fold.

4.6 Single cell amplification with MALBAC

A similar device was developed by Yu et al. where they used the recently developed off-chip MALBAC as a WGA step on mouse embryonic stem cells (42). This device was designed to minimize the contamination risk, preparation time and manual expertise while still performing a full DNA amplification which is beneficial for medical and forensic cases. The downstream process after amplification with the microfluidic device was to analyse copy number variants on the genome.

The compartments and channels on the chip were made of PDMS and were bonded with a silicon waver. As the MALBAC used thermocycling for the amplification, silicon was chosen because of the good heat conductance, which is regulated by a TE device. The device contained 8 separate reaction units which contained three separate chambers, with a height of 50μm, which are controlled by micro valves which close manually (Figure 11). Above the device a mono-colour camera with microscopic abilities was placed to monitor the cell during the process and to decide when to close or open micro valves. The distribution channels of the device had a common width of 250μm, while the chambers itself were separated by 420μm channels.

Figure 10. The specially designed micropillars on the device with in yellow the trapped DNA after cell lysis. The other components are washed away and indicated with blue (40).

(17)

16 The suspension of cells was driven on a low flow level through the first channel where single cells were isolated manually with micro valves and microscopic monitoring. Of the 8 separate reaction units, they loaded 6 of them with cells and retained the other 2 for negative control of the device, total loading time was about 5 minutes. Cell lysis was performed in the first chamber (75nL) with a lysis buffer (30 mM Tris-HCl, 10 mM KCl, 5 mM EDTA, 0.5% Triton-X100, with 2 mg/mL Protease) for 90 minutes at 50°C. After this the protease lysate was deactivated for 20 minutes at 75°C.

After lysis, the first two chambers were filled with a preamplification buffer and mixed passively with the cell lysate. The MALBAC Single Cell WGA Kit from Yikon Genomice was used for MALBAC preamplification and amplification. As MALBAC needs thermocycling for the amplification and correct use of the kit, several heating and cooling steps were taken to execute this. During the preamplification 10 cycles were ran which each took 6 minutes each (for each cycle: 30 s at 20°C, 30 s at 30°C, 30 s at 40°C, 30 s at 50°C, 30 s at 60°C, 180 s at 70°C, 20 s at 95°C, and 10 s at 58°C). The MALBAC amplification was performed after loading the PCR buffer into all the chambers and ran by 16 thermal cycles for which took 4 minutes each (for each cycle: 20 s at 95°C, 10 s at 58°C, and 180 s at 70°C). After this final step, the amplicons were collected with a pipet at the end of the amplification unit. Between each step the common channel was rinsed with water and flushed with air at 4°C.

The WGA amplification resulted in an 8000-fold yield of the DNA from a single cell, where each chamber produced a total amount of 50ng usable DNA for downstream analysis. Compared with off-chip MALBAC the reagents needed are reduced to 1/30 due to the size of the chambers and the isolation of a single cell. This method was validated by comparison with off-chip quantitative PCR amplification products and found to be easily reproducible.

Figure 11. The chip used for MALBAC shown on the left side, the PCR chamber represents the part where MALBAC was performed. On the right an overview of the external heating and cooling installation, the microfluidic device was placed on top of it and monitored with the microscope above the whole setup (42).

(18)

17

5. Discussion and conclusions

These studied papers all describe successful amplification of single cells on microfluidic devices which are mainly used for medical purposes. An overview of the studied microfluidic devices and their characteristics show several comparisons but also notable differences (Table 1).

5.1 Chip characteristics

PDMS was used in all cases as chip material due to the low costs and easy fabrication. The low thermal conductivity has also beneficial effects as all studies used off-chip heating with incubation on hot plates. Three of the devices used glass as an addition on the chip, mostly for firmness and visual properties. As in all these studies in between monitoring of the cell trapping was performed with microscopes, the transparent properties of the PDMS were of high importance. In the studies of Li et al. (39), Yang et al (36), Li et al (29). and, Tian et al. (40), fluorescent labelling was used to detect different cell types whereby PDMS materials are preferred. This is probably why silicon only used for the device of Li et al. as a bottom layer as the transparency is lower (39).

The precoating with BSA was studied by Yang et al. and Chu et al. (36,37), which should be beneficial for amplification due to the reduction of inhibitory components. As shown no differences in amplification yield nor time were reported between the studies with and without the BSA treatment. The BSA solution has proven to be of value in PCR reactions but show no difference in MDA on microfluidic devices.

The different compositional characteristics of the devices did not result in different success rates of the amplification reactions. The devices in which lysis and amplification reactions are performed in separate chambers which are locked with microvalves could be less sensitive for DNA loss during the reactions and therefore more preferred in forensic case work. Although the DNA strands are captured on micro pillars in the study of Tian et al., the continuous flow could result in DNA loss before amplification starts (40).

Especially for forensic purposes, the specificity of the DNA analysis should be as high as possible which could be achieved with a controlled device. The manufacturing of these devices is more complicated due to the implementation of the valves but required for DNA analysis from mixtures found on crime scenes. As the study of Li et al. (2019) showed, with controlled devices an equal error rate could be achieved compared to off-chip MDA (29).

Another difference between the devices is the number of separate chambers and channels to perform the single cell DNA amplification in. The device of Chu et al. with the SISSORS concept is only to amplify one single cell on the whole chip while the chip of Li et al. (2018), was able to perform 20 amplifications simultaneously (37,39). In forensic case work, the possibility to amplify several cells at the same time is preferred due to the DNA mixtures sampled from a scene. This option provides the investigators to amplify 20 single cells from a swab instead of just one which reduces time and costs.

As the MALBAC device showed similar amplification yield and time compared to the other studies, the method is more complex, due to thermal cycling, and requires more external equipment such as heating (42). The design of the chip could be used for MDA reactions as it is completely made of PDMS and consists of several chambers separated by micro valves. The device can amplify 8 simultaneous amplification reactions and with PDMS thermal heating could be easily performed, which could be useful in forensic case work

(19)

17

Table 1. Overview of microfluidic devices and amplification characteristics Amplification

Method material Chip Amplification time / temperature Total time of the device Kit used Amplification yield DNA amount as input Name, Year and Ref.

MDA PDMS 3 hours / 30°C 3 hours and 6 min REPLI-g Single Cell ≥ 250pg Single cell / ~6pg 2018 (39) Li et al.

MDA PDMS/Glass 2 hours / 34°C 3 hours and 20 min GenomiPhi V2 Illustra ~8.27ng Single cell / ~6pg Yang et al. 2014 (36)

MDA PDMS 16 hours/ 32°C 16 hours and 13 min REPLI-g Single Cell 30-50ng Single cell / ~6pg 2019 (29) Li et al.

MDA PDMS/Glass 15 hours / 30°C 15 hours and 20 min Self-designed 36ng Single cell / ~6pg Chu et al. 2017 (37)

MDA PDSM/Glass 3.5 hours / 33°C 4 hours and 30 min UltraFast mini REPLI-g 60ng Single cell / ~6pg Tian et al. 2018 (40)

(20)

18

5.2 Amplification yield

As one of the important purposes of WGA in forensic context, the amplification yield of these different devices should be addressed. All devices used single cells as an input for the amplification reactions which contain ~6-7pg of double stranded DNA as the input for STR-PCR off-chip is 100pg at least (4). A 25-fold amplification is the minimal yield that these devices should deliver, but more if possible. The devices discussed in this study all reach this requirement. They were all able to amplify the DNA from the single cell to the 100pg-requirement. As Chu et al. and Tian et al. both denatured the double stranded DNA into single stranded DNA before the amplification started resulted in comparable amplification yields (37,40). Another notable conclusion could be drawn as the amplification yield is closely link to the amplification time, longer amplification time results in higher DNA yield.

There were no notable differences between the different commercially available amplification kits that were used on these devices. The REPLI-g kits from Qiagen and Illustra GenomiPhi were both able to amplify single cell DNA on microfluidic levels. The study of Tian et al., showed the highest amplification yield which resulted in 60ng of amplicons while using the REPLI-g UltraFast mini kit (40).

A different study showed similar results for MDA in microdroplets on a high throughput workflow (43). In this study on 129 mammary epithelial cells simultaneously, the obtained ~60ng amplified DNA from single cells and >80% genome coverage. This shows that MDA performs equally in microdroplets as in chamber-controlled and flow-through devices. The droplet MDA system used here requires several in between steps, which could be performed manual or automatically and MDA reactions were performed for 18-20 hours. Although these microdroplets have shown their potential, time is the limiting factor to use this device in forensic context.

5.3 Amplification time

Another important requirement of the microfluidic devices is the overall amplification time as in several forensic cases the time to maintain a DNA profile could be crucial in the police investigation. These studies show two different categories in amplification time i) around 4 hours or ii) around 16 hours. The differences could be explained by number of steps prior to the amplification e.g. separation and lysis. As the study of Yang et al. and Tian et al. both had a preparation that took longer than one hour, the other four demonstrated to reduce this time to even 6 minutes (36,40). The difference could be explained by the lysis buffer that was used as they require different periods of execution. Also, additional washing steps were performed by the study of Tian et al. which results in a one-hour preparation before MDA could be performed.

Another explanation for this time difference is could be the time MDA itself is performed. As in the studies of Li et al. and Chu et al. (29,37) the amplification time is about 15 and 16 hours respectively the amplification time of the device of Yang et al.(36) was only 2 hours. This difference could be explained by the off-chip MDA protocol which requires about 15 hours of amplification while the microfluidic devices are a scaled down version of the conventional laboratory techniques.

The amplification time and amplification yield are shown to be related in these studies. As the methods of Li et al. and Chu et al. showed higher amplification yields than the methods of Li et al. and Yang et al., their amplification time is five times as long (29,36,37,40). As both are important factors for the evaluation for forensic purposes, a trade-off should be made at the point in the investigation. If time plays an important role, the more rapid devices should be preferred as in later stadia of the investigation, the methods with higher DNA yield could be used.

(21)

19

5.4 Recommendations for forensic application

As the above-mentioned microfluidic devices are all able to perform successful amplification of the whole genome from a single cell, several characteristics are more suitable for forensic implementation then others. Time plays an important role in many police investigations, so a short timeframe is preferred and most suitable in case work. The devices that could perform MDA within the 4 hours are studied by Li et al. , Yang et al. and Tian et al. (29,30,32). While these amplification numbers show great potential, the device's design shows less precision in cell trapping as it is a flow-through design without the control of chambers and valves. As a recommendation, PDMS should be used as chip material as it is low in costs and easy to fabricate which is preferred in forensic case work whereby chips are disposed to prevent contamination.

As a recommendation, the design of Li et al. which is able to perform 20 simultaneously MDA reactions on a complete PDMS chip at approximately 3 hours could be evaluated on forensic case work(39). Although this study did not report in between DNA quantification, the implementation on sequencing resulted successfully. For future research, this design could be studied in more detail to determine the precision and direct amplification yield with this device. In combination with the REPLI-g single cell amplification kit, or the other kits, this design and method could be evaluated for forensic applications. Preferably the MDA kit GenomiPhi™ and PCR-kit GenomePlex™ as they both showed bias decreasing potential (2). The amplification bias is not reported in all of the studies but plays an important role in forensic cases, therefore future studies should include these numbers and evaluate the validity of the WGA on microfluidic devices.

Overall, the implementation of on-chip WGA before off-chip STR-PCR is a promising addition to the forensic DNA analysis and profiling. These chips have shown that MDA can be performed at the micro level and that single cells can be amplified with great success. The minimum DNA input for full STR-PCR amplification and profiling could be achieved if these devices are included in the DNA analysis workflow. For further development, these chips could be used on site for DNA sampling whereby it reduces the risk of contamination. Since MDA on the chip could only amplifies a single cell, the DNA profile distinctiveness could be increased, helping the criminal justice system with their convictions. As concluded in off-chip MDA before conventional STR-PCR, even in complex mixtures, the amount of DNA from a minor contributor increases which results in more inclusive evidence (23).

(22)

20

6. References

1. Mapes AA, Kloosterman AD, van Marion V, de Poot CJ. Knowledge on DNA Success Rates to Optimize the

DNA Analysis Process: From Crime Scene to Laboratory. J Forensic Sci. 2016;

2. Ballantyne KN, van Oorschot RAH, Mitchell RJ. Comparison of two whole genome amplification methods

for STR genotyping of LCN and degraded DNA samples. Forensic Sci Int. 2007;

3. Gill P, Whitaker J, Flaxman C, Brown N, Buckleton J. An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA. Forensic Sci Int. 2000;

4. Cho S, Shin KJ, Bae SJ, Kwon YL, Lee SD. Improved STR analysis of degraded DNA from human skeletal

remains through in-house MPS-STR panel. Electrophoresis. 2020;

5. Romsos EL, Vallone PM. Rapid PCR of STR markers: Applications to human identification. Forensic Science

International: Genetics. 2015.

6. Verheij S, Harteveld J, Sijen T. A protocol for direct and rapid multiplex PCR amplification on forensically relevant samples. Forensic Sci Int Genet. 2012;

7. Maciejewska A, Jakubowska J, Pawłowski R. Whole genome amplification of degraded and nondegraded DNA

for forensic purposes. Int J Legal Med. 2013;

8. Ambers A, Turnbough M, Benjamin R, Gill-King H, King J, Sajantila A, et al. Modified DOP-PCR for improved STR typing of degraded DNA from human skeletal remains and bloodstains. Leg Med. 2016; 9. Bruijns B, van Asten A, Tiggelaar R, Gardeniers H. Microfluidic devices for forensic DNA analysis: A review.

Biosensors. 2016.

10. Bruijns BB, Costantini F, Lovecchio N, Tiggelaar RM, Di Timoteo G, Nascetti A, et al. On-chip real-time

monitoring of multiple displacement amplification of DNA. Sensors Actuators, B Chem. 2019;

11. Deleye L, Tilleman L, Van Der Plaetsen AS, Cornelis S, Deforce D, Van Nieuwerburgh F. Performance of

four modern whole genome amplification methods for copy number variant detection in single cells. Sci Rep. 2017;

12. Alberts B, Johnson A, Lewis J, Morgan D, Raff M, Roberts K, et al. Molecular Biology of the Cell. Molecular Biology of the Cell. 2017.

13. Chen M, Song P, Zou D, Hu X, Zhao S, Gao S, et al. Comparison of Multiple Displacement Amplification

(MDA) and Multiple Annealing and Looping-Based Amplification Cycles (MALBAC) in single-cell sequencing. PLoS One. 2014;

14. Börgstrom E, Paterlini M, Mold JE, Frisen J, Lundeberg J. Comparison of whole genome amplification techniques for human single cell exome sequencing. PLoS One. 2017;

15. Huang L, Ma F, Chapman A, Lu S, Xie XS. Single-Cell Whole-Genome Amplification and Sequencing: Methodology and Applications. Annu Rev Genomics Hum Genet. 2015;

16. Cheung VG, Nelson SF. Whole genome amplification using a degenerate oligonucleotide primer allows hundreds of genotypes to be performed on less than one nanogram of genomic DNA. Proc Natl Acad Sci U S A. 1996;

17. Lasken RS. Single-cell genomic sequencing using Multiple Displacement Amplification. Current Opinion in

Microbiology. 2007.

18. Hansen KD, Brenner SE, Dudoit S. Biases in Illumina transcriptome sequencing caused by random hexamer

priming. Nucleic Acids Res. 2010;

19. Zong C, Lu S, Chapman AR, Xie XS. Genome-wide detection of single-nucleotide and copy-number variations of a single human cell. Science (80- ). 2012;

20. Lasken RS. Single-cell sequencing in its prime. Nat Biotechnol. 2013;

21. Gopinath S, Zhong C, Nguyen V, Ge J, Lagacé RE, Short ML, et al. Developmental validation of the Yfiler®

Plus PCR Amplification Kit: An enhanced Y-STR multiplex for casework and database applications. Forensic Sci Int Genet. 2016;

22. Wang DY, Chang CW, Hennessy LK. Rapid STR analysis of single source DNA samples in 2 h. Forensic Sci

Int Genet Suppl Ser. 2009;

23. Zhu Q, Fang T, Zhou Y, Yang Y, Cao Y, Wang Q, et al. Effect of phi29 polymerase-based multiple strand

displacement whole genome amplification on the proportion in DNA mixtures. Forensic Sci Int Genet Suppl Ser. 2019;

24. Rakay CA, Bregu J, Grgicak CM. Maximizing allele detection: Effects of analytical threshold and DNA levels on rates of allele and locus drop-out. In: Forensic Science International: Genetics. 2012.

25. Easley CJ, Karlinsey JM, Bienvenue JM, Legendre LA, Roper MG, Feldman SH, et al. A fully integrated microfluidic genetic analysis system with sample-in-answer-out capability. Proc Natl Acad Sci U S A. 2006;

26. DuVall JA, Le Roux D, Thompson BL, Birch C, Nelson DA, Li J, et al. Rapid multiplex DNA amplification

on an inexpensive microdevice for human identification via short tandem repeat analysis. Anal Chim Acta. 2017;

27. Han J, Gan W, Zhuang B, Sun J, Zhao L, Ye J, et al. A fully integrated microchip system for automated forensic short tandem repeat analysis. Analyst. 2017;

(23)

21

28. Bruijns B, Tiggelaar R, Gardeniers H. A microfluidic approach for biosensing DNA within forensics. Appl

Sci. 2020;

29. Li R, Jia F, Zhang W, Shi F, Fang Z, Zhao H, et al. Device for whole genome sequencing single circulating tumor cells from whole blood. Lab Chip. 2019;

30. Nielsen JB, Hanson RL, Almughamsi HM, Pang C, Fish TR, Woolley AT. Microfluidics: innovations in materials and their fabrication and functionalization. Analytical Chemistry. 2020.

31. Nge PN, Rogers CI, Woolley AT. Advances in microfluidic materials, functions, integration, and applications. Chemical Reviews. 2013.

32. Bhagat AAS, Jothimuthu P, Papautsky I. Photodefinable polydimethylsiloxane (PDMS) for rapid

lab-on-a-chip prototyping. Lab Chip. 2007;

33. Vigolo D, Rusconi R, Piazzaa R, Stone HA. A portable device for temperature control along microchannels.

Lab Chip. 2010;

34. Miralles V, Huerre A, Malloggi F, Jullien M-C. A Review of Heating and Temperature Control in Microfluidic Systems: Techniques and Applications. Diagnostics. 2013;

35. Sia SK, Whitesides GM. Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies. Electrophoresis. 2003.

36. Yang Y, Swennenhuis JF, Rho HS, Le Gac S, Terstappen LWMM. Parallel single cancer cell whole genome

amplification using button-valve assisted mixing in nanoliter chambers. PLoS One. 2014;

37. Chu WK, Edge P, Lee HS, Bansal V, Bafna V, Huang X, et al. Ultraaccurate genome sequencing and haplotyping of single human cells. Proc Natl Acad Sci U S A. 2017;

38. Veltkamp HW, Monteiro FA, Sanders R, Wiegerink R, Lötters J. Disposable DNA amplification chips with

integrated low-cost heaters. Micromachines. 2020;

39. Li R, Zhou M, Yue C, Zhang W, Ma Y, Peng H, et al. Multiple single cell screening and DNA MDA amplification chip for oncogenic mutation profiling. Lab Chip. 2018;

40. Tian HC, Benitez JJ, Craighead HG. Single cell on-chip whole genome amplification via micropillar arrays for reduced amplification bias. PLoS One. 2018;

41. Rahbari R, Sheahan T, Modes V, Collier P, Macfarlane C, Badge RM. A novel L1 retrotransposon marker for

HeLa cell line identification. Biotechniques. 2009;

42. Yu Z, Lu S, Huang Y. Microfluidic whole genome amplification device for single cell sequencing. Anal Chem. 2014;

43. Leung K, Klaus A, Lin BK, Laks E, Biele J, Lai D, et al. Robust high-performance nanoliter-volume single-cell multiple displacement amplification on planar substrates. Proc Natl Acad Sci U S A. 2016;

44. Yang Y, Rho HS, Stevens M, Tibbe AGJ, Gardeniers H, Terstappen LWMM. Microfluidic device for DNA

(24)

22

7. Appendix A - Search strategy

1) Start with the given literature from the description 🡪 2 a. With this literature found five more articles 🡪 5 2) Used a study book 🡪 1

3) Literature search with https://scholar.google.com/

Scholar, articles since 2016. Scanned these articles by title, used the forensic relevant ones 🡪 3

Based on the references from these three articles, found five more 🡪 5 Scholar, articles since 2016. To many so used more specifics.

Scholar, articles since 2016. Used this article 🡪 1

Based on the references of this article, found two more 🡪 2

Scholar, articles since 2016. Scanned the articles by title and first sentences, used five relevant 🡪 5

Based on the references of these articles, found six more 🡪 6

Scholar, articles since 2019. Scanned the articles by title and summary, used two relevant 🡪 2

Based on the references of this article, found one more 🡪 1

Based on the references of this article, found one more 🡪 1 Terms: [forensics AND wga] 196 articles

Terms: [forensics AND whole genome amplification] 11400 articles

Terms: [allintitle: forensics AND whole genome amplification] 1 article

Terms: [microfluidic devices AND multiple displacement amplification] 15,400 articles

Terms: [microfluidic devices AND forensics] 2090 articles

Terms: [MDA AND STR] 3180 articles

Terms: [rapid AND STR AND microfluidic AND forensics] 184 articles Terms: [microfluidic devices AND materials AND WGA] 395 articles

(25)

23 Scholar, articles since 2019. Scanned the articles by title and summary, relevant articles already included

Scholar, articles since 2019. Scanned the articles by title and summary, relevant articles already included

4) All these together sum up to 44 articles that are used for this literature review. Terms: [microfluid devices AND forensics AND MDA] 126 articles

Single cell DNA-amplification on-chip