Skip to main content

Nanofluidic Methods for Mapping Epigenetic and Genomic Variation

Final Report Summary - DNAMAP (Nanofluidic Methods for Mapping Epigenetic and Genomic Variation)

The project set out to advance the fast detection of methilcillin-resistant Staphylococcus aureus (MRSA) strains. These ubiquitous strains of bacteria are a threat to patients with a weak immune system. MRSA strains have evolved as a consequence of the wide application of broad-band antibiotics during the past decades. MRSA in patients admitted to hospitals needs to be identified in order to provide special treatment for these patients. Diagnosis needs to be high throughput, i.e. rapid, and inexpensive. Within DNAMAP, identification of MRSA was set out to happen by optical mapping of partially stained genomic DNA.

The project was based on micro- and nanofluidic systems in which the genomic DNA could be introduced, manipulated, and imaged. The imaged DNA could then be mapped to an existing reference genome, and inserts which evoke antibiotic resistance were to be identified. The aim of the work was to show how micro- and nanofluidic systems can be used for the mapping procedure, and to show the system's potential for automated operation.

The work performed consisted of designing, fabricating, characterizing, and operating the fluidic devices which were fabricated in fused silica. These devices contain nanochannels that are 150 nm in width and height. They are used to force-stretch DNA to approximately 50% of its contour length. Images of stretched DNA, usually no longer than 200 μm in contour length, were then mapped to an existing reference genome. The recorded data were however not sufficient to doubtlessly assign the imaged stretches of DNA to specific regions on the reference genome. Most importantly, this was due to the algorithm for mapping DNA stretches. This algorithm had been established prior to the start of the DNAMAP project, but inconsistencies within the algorithm have been found which eventually questioned further efforts of imaging and mapping small DNA stretches.

Instead, the continued work focused on using a newer, more promising model for DNA mapping. That model relied on significantly larger stretches of DNA (more than 475 μm, vs. the previous 200 μm). An additional significant improvement to the previously used device consisted of the ability to stretch the DNA to 100 % of its contour length, yielding the best possible imaging resolution. DNA samples were selected according to their contents of long and seemingly disentangled strands. Initial work on such samples where these strands were stretched and immobilized showed useful structural patterns on the strands, and therefore considerable effort was invested to master the micro/nanofluidic system that was used to manipulate the DNA strands in order to stretch them. This fluidic system was comprised of a wide, cross-shaped slit in which the DNA could be stretched by shear-flow. Microchannels were used to interface this slit and for funneling DNA into it. It was found that long DNA molecules which were funneled into the nanoslit did however not show any structural patterns. The discrepancy between the observation of such patterns when the molecules were immobilized and when the molecules were stretched in a nanofluidic system was investigated in depth. It was concluded that these long DNA strands which are seemingly not entangled, are probably agglomerations of two or few strands. The reason that single strands could be seen when they were immobilized was attributed to the fact that many strands can be screened this way. In the fluidic device however, there is a bias towards selecting entangled molecules since they are the brightest. The chance to select molecules that are not entangled was therefore very small.

When this was realized, an attempt to reduce the entanglement was undertaken. At this point, the collaborator who provided the DNA samples withdrew his support which finally led to the end of mapping MRSA DNA with a cross-slit device. The decision to withdraw the support was on one hand a consequence of the long time the project had been running, and despite giving promising preliminary results, did not lead to a break-through. On the other hand, this work was motivated by a potential great reduction in screening cost for MRSA. Over the time that this project ran, alternative technologies have emerged which offered a reduction in screening cost and made the DNA mapping work in DNAMAP less attractive to be pursued further.

Alternative avenues were pursued in micro- and nanofabrication, including methods of characterization and device finishing. This was an integral part of the training programme in DNAMAP. New competences in cleanroom fabrication have been acquired, most importantly dry and wet etching techniques, electron-beam lithography, and device bonding. As a part of that, characterization abilities were mainly extended to Scanning Electron Microscopy (SEM). For such complex cleanroom processes, there was need to apply skills in device planning and process design.

The programme also encompassed teaching and student supervision. In collaboration with students in the work group, device designs and fabrication processes were developed. This was for instance the case for designing pinch-flow devices as well as inlet structures for disentangling DNA strands that had been isolated from single cells.

Another investigation dealt with the behaviour of short lambda-DNA molecules which enter nanochannels as small as 50 nm wide and tall. Such channels were fabricated with electron-beam lithography in fused silica nanofluidic devices. Unlike 150 nm channels used earlier which only allow DNA to be stretched to 50 % of its contour length, 50 nm channels allow DNA stretching to 85% of its contour length, hence, improved imaging resolution ensues. Apart from being a challenge in fabrication, small channels in fluidic devices entail high hydraulic resistance and large energy barriers when a DNA molecule needs to be transferred into such a channel. The large hydraulic resistance entails diminished flow of liquid to pull DNA molecules into the channels. It was found that smart inlet structures provide a stepping stone for DNA molecules to transit from a microchannel where they are present in random coils towards a stretched conformation in nanochannels. (c.f. figure below) The passing of DNA molecules was facilitated by a range of electric potentials applied across the nanochannels. When an intermediate slit structure was introduced between the nanochannels and the microchannels from which DNA molecules were supplied, the rate of DNA molecules passing through the nanochannels could be significantly improved.