4
$\begingroup$

Just wondering whether anyone can point me to a research paper which describes how the scoring values are generated by the Bruker software for species identification of bacteria in MALDI-TOF MS. For example, there are countless papers describing the Bruker output score as green (species-level confidence), yellow (genus-level), or red (neither species- nor genus-level confidence). However, none of the papers that I see are discussing how these numbers are actually calculated within the Bruker system. Even the Bruker brochure glosses over the point: "After the acquisition of the spectral data has been completed, a report is generated. The result for each sample is clearly listed under ‘Organism (best match)’ accompanied by the resulting score and an appropriate ‘traffic light’ color scheme." Can anyone point me in the direction of a paper or at least some online information such as wikipedia describing how Bruker came up with these numbers? It's not just an arbitrary number, right?

Another generic description of the scores is in this presentation: "Higher the log (score), higher the similarity between mass spectrum of isolate & the database entry in the reference library."

$\endgroup$
3
  • $\begingroup$ From the 5th page of the brochure: "Unknowns are then compared to the MSP library using a superior pattern-matching approach. This includes peak positions and intensities, ensuring the highest possible levels of accuracy and reproducibility across the complete range of microorganisms". It seems like it's their secret ingredient and they're not really keen in sharing it with the world :/ $\endgroup$
    – Álvaro
    May 24, 2021 at 22:59
  • $\begingroup$ This paper as well as this other paper both cite this other paper in their methods section when they talk about the parameter settings for the MALDI Biotyper, and the paper itself doesn't give much detail. That's all I've found so far, and I'm probably showing you something you've seen already $\endgroup$
    – Álvaro
    May 24, 2021 at 23:02
  • $\begingroup$ "For identification scoring, the software correlates signal intensities of matched signals of mass spectra. The three scores obtained from such a procedure are multiplied and normalized to a value of 1,000 and its common logarithm. Log scores over 2 are considered ... whereas log scores over 1.7 generally ... Log scores of 3 are obtained when spectra are matched with themselves." Sauer, S., Freiwald, A., Maier, T., Kube, M., Reinhardt, R., .., & Geider, K. (2008). Classification and identification of bacteria by mass spectrometry and computational analysis. PloS one, 3(7), e2843. $\endgroup$
    – There
    May 25, 2021 at 17:47

1 Answer 1

1
$\begingroup$

For identification scoring, the software correlates signal intensities of matched signals of mass spectra. The three scores obtained from such a procedure are multiplied and normalized to a value of 1,000 and its common logarithm. Log scores over 2 are considered as reliable identification of a bacterial species (Table 4), whereas log scores over 1.7 generally indicate reliable identification of bacterial genera. Log scores of 3 are obtained when spectra are matched with themselves.

Sauer, S., Freiwald, A., Maier, T., Kube, M., Reinhardt, R., Kostrzewa, M., & Geider, K. (2008). Classification and identification of bacteria by mass spectrometry and computational analysis. PloS one, 3(7), e2843.

Going further, from a random thread on ResearchGate:

https://www.researchgate.net/post/Does_anyone_know_how_is_the_p-value_calculated_by_Bruker_Biotyper_during_MALDI-MS_fingerprinting_of_bacteria

Number of peaks matched in your unknown sample divided by the number of peaks in the unknown sample, this is given a score (out of 10)

Number of peaks matched in your unknown sample divided by the number of peaks in the reference sample, this is given a score (out of 10)

Score of relative intensities of matching peaks (out of 10)

Total score out of 1000

Log10 score out of 3

There is no relationship of the score with a p-value nor you can calculate the statistics because the bruker database is not open and you can not know the values for what the score was calculated to do the statistical analysis

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.