Correcting biological infrared spectroscopy data for atmospheric gases and Mie scattering
(2020) FYTM04 20201Computational Biology and Biological Physics - Has been reorganised
- Abstract
- Infrared absorption microscopy is a powerful chemometric tool with a wide variety of applications. It is, however, subject to considerable disturbances from atmospheric gases and scattering effects. As experimental fixes are not always applicable or advisable, data therefore needs to be computationally corrected before any attempt at an interpretation. This work designed, improved and validated several methods to correct such imaging data with varying success. It also provided a reliable way to assess the efficacy of such correction, as well as a set of methods to produce synthetic test data.
- Popular Abstract
- Biochemical screening techniques (such as cancer diagnosis, water testing, etc.) used today tend to be unreliable, slow, and most importantly expensive. Infrared spectroscopic imaging has the potential to make this process easier, faster, and cheaper. It is, however, subject to disturbances that make it very difficult to interpret. Techniques to get rid of these disturbances exists, but have the drawback of making the whole process either slower, more expensive, or both. Solving the problem with an algorithm, however, would allow to get rid of the drawbacks while keeping the advantages of the technique.
Molecules have long been known to vibrate at specific frequencies when excited by electromagnetic radiation, in a similar way as a... (More) - Biochemical screening techniques (such as cancer diagnosis, water testing, etc.) used today tend to be unreliable, slow, and most importantly expensive. Infrared spectroscopic imaging has the potential to make this process easier, faster, and cheaper. It is, however, subject to disturbances that make it very difficult to interpret. Techniques to get rid of these disturbances exists, but have the drawback of making the whole process either slower, more expensive, or both. Solving the problem with an algorithm, however, would allow to get rid of the drawbacks while keeping the advantages of the technique.
Molecules have long been known to vibrate at specific frequencies when excited by electromagnetic radiation, in a similar way as a guitar string emits a note when picked by the guitarist’s finger. By recording how a sample reacts to this excitation, scientists can determine its composition in much the same way as the musician can hear the notes that make up a chord. In real life, however, samples are not as simple as the sound emitted by a carefully played 6-string guitar in a quiet room, but would be better compared to a piano sonata being played in the middle of Times Square at rush hour. Numerous and hard to predict vibrations appear and overlap with the fine melody. So much so that it is difficult to identify what corresponds to the actual note being played and what is background noise. In Times Square, the sound of cars and people passing by is intense. In a biological sample, this “noise” can come from a lot of things, but originates mainly from the water vapor and carbon dioxide which surround it and which also vibrate in those frequencies of interest. It is of course possible to isolate the sample from much of these perturbations by keeping it in a very controlled environment, but this requires additional time and equipment. Nobody wants to (or can) rent a recording studio every time they want to listen to music. This is why the most promising way to get rid of the unwanted vibrations is instead to remove them with the use of a computer algorithm. Indeed, even though the background noise is unpredictable, it still has some key differences from the music which make it possible—to some extent—to separate them. We therefore could, in theory, “listen in” to the vibrations emitted by, for example, cancer cells markers in the middle of the chaos of a biopsy. Then, the equipment needed to produce this kind of diagnosis would fit in a backpack and cost thousands of dollars, instead of complete labs worth millions today. It would be almost instant, and very easily deployed in almost any circumstances, and not just in a hospital. Additionally, this technique would not be restricted to cancer screening, and not even to medical applications. Being able to identify the precise composition of a sample is useful in numerous fields and could help to, for instance, check the drinkability of a water source, or understand how certain fungi can break down plastics. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9024329
- author
- Pissot, Stéphan LU
- supervisor
-
- Carl Troein LU
- organization
- course
- FYTM04 20201
- year
- 2020
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- FTIR, spectroscopy, chemometrics, infrared absorption microscopy, atmospheric correction, scattering correction
- language
- English
- id
- 9024329
- date added to LUP
- 2020-07-14 10:37:08
- date last changed
- 2021-01-25 08:30:28
@misc{9024329, abstract = {{Infrared absorption microscopy is a powerful chemometric tool with a wide variety of applications. It is, however, subject to considerable disturbances from atmospheric gases and scattering effects. As experimental fixes are not always applicable or advisable, data therefore needs to be computationally corrected before any attempt at an interpretation. This work designed, improved and validated several methods to correct such imaging data with varying success. It also provided a reliable way to assess the efficacy of such correction, as well as a set of methods to produce synthetic test data.}}, author = {{Pissot, Stéphan}}, language = {{eng}}, note = {{Student Paper}}, title = {{Correcting biological infrared spectroscopy data for atmospheric gases and Mie scattering}}, year = {{2020}}, }