Figure 3: Diffusion-like local mapping algorithm for detecting target bead with real-world location on a chip. | Nature Communications

Figure 3: Diffusion-like local mapping algorithm for detecting target bead with real-world location on a chip.

From: A high-throughput optomechanical retrieval method for sequence-verified clonal DNA from the NGS platform

Figure 3

(a) Schematic diagram of our mapping algorithm. The 454 Junior normally offers ~105 sequences with accordant pixel positions of CCD. From two arbitrary reference points, the corresponding sequence-labelled well location can be determined by adjusting for the scale and rotational angle of the sequencing pixel domain and overlap. (b) However, due to random and non-linear distortion of the sequencer’s imaging system, one-step global transformation leads to locational error. Approximately 20% of the pixels are mapped in a false position that is not distinguishable, severely reducing the reliability of all of the location data (Supplementary Fig. 22). Yellow flags indicate the reference points of each mapping calculation. A colour bar shows the pixel-wise distance between the mapped pixels and accordant well centre. The threshold value is ~13.5 pixels. (c) We reduced the effect of imaging distortion to a negligible level by localizing the region of interest. The whole-chip area was divided into 300 subdomains with a slight overlap. (d) One subdomain completes the location mapping by supplying two new reference points to the adjacent subdomain. Local mapping propagates from the initial matched subdomains throughout the whole chip. (e) Twenty-four beads were retrieved from eight evenly distributed regions to verify the local mapping algorithm. The left side of the figure describes the target well location on the stitched chip image and the right side of the figure shows the correct retrieval results. (f) Retrieved beads were amplified and identified with the Sanger method, showing matched results for each of the reference sequences.

Back to article page