Note: Descriptions are shown in the official language in which they were submitted.
CA 02490509 2012-12-19
IMPROVED METHOD OF REDUCING THE EFFECTS OF BACKGROUND NOISE IN
MASS SPECTRA
The present invention relates to a mass spectrometer
and a method of mass spectrometry.
Background chemical noise in a mass spectrum can be
particularly problematic. The background chemical noise
observed in mass spectra often has a periodic nature
especially at mass to charge ratios less than 1000. As
will be understood by those skilled in the art, all
elements have near integral masses. Carbon-only graphite
has, by definition, an exact integer mass of 12 and all
other molecules of the same nominal mass will have an
exact mass which is not quite an exact integer value but
yet which is only slightly higher or lower than the
corresponding mass of carbon-only graphite.
The most mass sufficient ions formed from organic
and biological molecules are saturated hydrocarbons and
the most mass deficient ions formed from organic and
biological molecules are saturated bromocarbons.
Saturated hydrocarbons have a mass sufficiency of about
0.1%. Accordingly, a saturated hydrocarbon with a
nominal mass of 100 will have an exact mass of about
100.1 and likewise a saturated hydrocarbon with a nominal
mass of 200 will have an exact mass of about 200.2.
Saturated bromocarbons have a mass deficiency of about
0.1%. Accordingly, a saturated bromocarbon with a
nominal mass of 100 will have an exact mass of about 99.9
and likewise a saturated bromocarbon with a nominal mass
of 200 will have an exact mass of about 199.8. As a
result, at a nominal mass of 200 singly charged ions can
be expected to have exact masses which fall within a
relatively narrow mass to charge ratio range of 199.8 to
CA 02490509 2004-12-17
-2-
200.2. Similarly, at a nominal mass of 201 singly
charged ions can be expected to have exact masses which
fall within a similar relatively narrow mass to charge
ratio range 200.8 to 201.2. It will therefore be
appreciated that no singly charged ions having exact
masses in the range 200.2 to 200.8 will be observed.
Accordingly, at relatively low mass to charge ratios the
chemical background noise in mass spectra (which is
predominantly singly charged) typically exhibits a
distinct periodicity of approximately 1 atomic mass units
(amu).
For singly charged ions having mass to charge ratios
of 500 or more, the range of forbidden exact masses
theoretically shrinks to zero and hence it might be
expected that the chemical background noise would no
longer exhibit a periodicity of approximately 1 atomic
mass unit. However, in practice, saturated hydrocarbons
and saturated bromocarbons are rarely encountered when
mass analysing biochemical samples such as proteins and
peptides. Accordingly, the chemical background noise in
mass spectra relating to biochemicals or biomolecules
commonly exhibits a distinct periodicity of approximately
1 atomic mass unit at mass to charge ratios in excess of
500. Indeed, mass spectra commonly exhibit a distinct
periodicity of approximately 1 atomic mass unit at mass
to charge ratios up to about 2000 and periodic background
noise may, in some circumstances, be observed at mass to
charge ratios in excess of 2000.
Most non-halogenated organic molecules have a mass
sufficiency in the range 0.0% to 0.1%. Therefore,
assuming that halogenated compounds are absent, then it
will be appreciated that the chemical background noise
can still be expected to have a periodicity of
CA 02490509 2004-12-17
- 3 -
approximately 1 atomic mass unit at mass to charge ratios
up to 1000. Indeed, in practice, chemical background
noise having a periodicity of approximately 1 atomic mass
unit is commonly observed when mass analysing ions
derived from biomolecules having mass to charge ratios up
to about 2000.
Many mass spectrometric techniques have detection
limits which are restricted or otherwise compromised by
the presence of chemical background noise. The precise
chemical nature of the background noise is often unknown
and the presence of unwanted chemical background noise
can adversely affect mass measurement accuracy especially
if an analyte signal is not fully resolved due to
chemical background noise.
Chemical background noise may, for example, arise
from impurities in solvents, analytes or reagents.
Impurities in drying or nebulizing gases can also cause
chemical background noise. Contamination of the solvent
or analyte delivery system or contamination within or on
the surfaces of an ionisation chamber can be a further
source of chemical background noise.
In Atmospheric Pressure Ionisation ("API") ion
sources such as Electrospray ("ESI"), Photo Ionisation
("APPI") or Atmospheric Chemical Ionisation ("APCI") ion
sources, chemical background can arise from the
clustering of solvent and analyte ions. In Chemical
Ionisation ("CI") ion sources chemical background can
arise from self-adduction of reagent gas ions or from
reagent gas contamination. In Matrix Assisted Laser
Desorption Ionisation ("MALDI") ion sources chemical
background can arise from matrix cluster ions.
In general the chemical background noise observed in
mass spectra tends to be complex in nature and may only
CA 02490509 2004-12-17
- 4 -
be partially mass resolved. The chemical background
noise tends to be singly charged and to have a periodic
nature with a repeat unit of approximately 1 atomic mass
unit. Amino acids have a mass sufficiency which varies
from about 1.00009 to about 1.00074, with a mean mass
sufficiency of approximately 1.00047. Accordingly,
biological samples commonly exhibit a periodicity of
approximately 1.0005 atomic mass units (Daltons).
A known approach to reducing the effects of periodic
background chemical noise in a mass spectrum is to
transform the mass spectrum into the frequency domain and
then to filter out noise components. Signals in the
transformed spectrum which are considered to represent
noise can then be removed at certain calculated
frequencies. An inverse transform is then applied to the
transformed spectrum in order to reproduce a mass
spectrum which exhibits reduced periodic background
noise.
Non-sinusoidal periodic noise will appear as a
series of sharp spikes and harmonics in the frequency
domain or transformed spectrum. Ion signals however,
since they are of relatively small extent in mass to
charge ratio, will tend to be smeared out across a
relatively broad range of frequencies. The different
characteristics of signal and noise in the frequency
domain or transformed spectrum can in theory at least be
used to allow the contribution of chemical background
noise in the overall spectrum to be reduced. However,
one problem with frequency domain filtering is that the
unprocessed time of flight mass spectra data will
comprise intensity data which is equally spaced in time
due to the acquisition electronics. Since flight time in
a Time of Flight mass analyser is proportional to the
CA 02490509 2004-12-17
- 5 -
square root of the mass to charge ratio of the ions, the
intensity data will be unequally spaced with respect to
mass to charge ratio. Accordingly, prior to filtering
the data in the frequency domain or transformed spectrum,
it is first necessary to process the mass spectral data
such that the intensity data is more equally spaced with
respect to mass to charge ratio. It is known to use an
interpolation algorithm to process the intensity data so
that the data becomes equally spaced with respect to mass
to charge ratio. However, disadvantageously, the use of
an interpolation algorithm significantly increases the
overall processing time.
In addition to increasing the overall processing
time, the known approach of reducing periodic noise in a
mass spectrum by filtering the data in the frequency
domain suffers from the problem that the application of a
filter to the frequency domain data to remove noise
components can actually result in additional noise and
discontinuities being present into the mass spectrum
after data in the frequency domain has been transformed
back into the mass to charge ratio domain. As a result,
artefacts or spurious peaks can appear in the final
processed mass spectrum which were not present in the
original mass spectral data.
Another problem with the known frequency domain
filtering approach is that a proportion of the desired
analyte signal will have frequency components which are
similar or identical to the frequency components
corresponding to unwanted background noise. Accordingly,
the removal of such components in the frequency domain
can lead to distortion both of the analyte ion peak shape
and also of the intensity of the analyte signal in the
final processed mass spectrum.
CA 02490509 2004-12-17
- 6 -
A yet further problem with the known frequency
domain filtering approach is in responding to changes in
the characteristic of the background noise as a function
of mass to charge ratio. The observed background noise
in a mass spectrum often takes on a different nature in
different portions of the mass spectrum i.e. the
background noise is often observed to vary as a function
of mass to charge ratio. If therefore a filter needs to
change shape as a function of mass to charge ratio in
response to the changing nature of the background noise,
then the mass spectrum must first be divided up into a
number of separate sections, each of which must then be
treated or filtered slightly differently. However,
discontinuities can then arise when a composite mass
spectrum is subsequently reconstructed from the separate
sections of data.
It is apparent therefore that the known frequency
domain filtering approach suffers from a number of
problems.
It is therefore desired to provide an improved
method of reducing the effects of background chemical
noise in mass spectra and in particular to reduce the
effects of background chemical noise having a periodic
nature.
According to the present invention there is provided
a method of mass spectrometry comprising:
determining an intensity distribution from a
plurality of different regions or portions of mass
spectral data or a mass spectrum;
estimating a background intensity for one or more
regions or portions of the mass spectral data or the mass
spectrum from the intensity distribution; and
CA 02490509 2004-12-17
, .
- 7 -
adjusting the intensity of one or more regions or
portions of the mass spectral data or the mass spectrum
in order to remove or reduce the effects of the estimated
background intensity.
The plurality of regions or portions of the mass
spectral data or the mass spectrum are preferably
discrete non-contiguous regions or portions. However,
according to less preferred embodiments the plurality of
regions or portions of the mass spectral data or the mass
spectrum may be substantially contiguous regions or
portions.
The plurality of regions or portions of the mass
spectral data or the mass spectrum preferably have a
periodicity selected from the group consisting of: (i) 0-
0.1 amu; (ii) 0.1-0.2 amu; (iii) 0.2-0.3 amu; (iv) 0.3-
0.4 amu; (v) 0.4-0.5 amu; (vi) 0.5-0.6 amu; (vii) 0.6-0.7
amu; (viii) 0.7-0.8 amu; (ix) 0.8-0.9 amu; (x) 0.9-1.0
amu; (xi) 1.0-1.1 amu; (xii) 1.1-1.2 amu; (xiii) 1.2-1.3
amu; (xiv) 1.3-1.4 amu; (xv) 1.4-1.5 amu; (xvi) 1.5-1.6
amu; (xvii) 1.6-1.7 amu; (xviii) 1.7-1.8 amu; (xix) 1.8-
1.9 amu; (xx) 1.9-2.0 amu; and (xxi) > 2.0 amu.
According to a preferred embodiment the plurality of
regions or portions of the mass spectral data or the mass
spectrum may have a periodicity of: (i) 0.4995-0.4996
amu; (ii) 0.4996-0.4997 amu; (iii) 0.4997-0.4998 amu;
(iv) 0.4998-0.4999 amu; (v) 0.4999-0.5000 amu; (vi)
0.5000-0.5001 amu; (vii) 0.5001-0.5002 amu; (viii)
0.5002-0.5003 amu; (ix) 0.5003-0.5004 amu; (x) 0.5004-
0.5005 amu; (xi) 0.9990-0.9991 amu; (xii) 0.9991-0.9992
amu; (xiii) 0.9992-0.9993 amu; (xiv) 0.9993-0.9994 amu;
(xv) 0.9994-0.9995 amu; (xvi) 0.9995-0.9996 amu; (xvii)
0.9996-0.9997 amu; (xviii) 0.9997-0.9998 amu; (xix)
0.9998-0.9999 amu; (xx) 0.9999-1.0000 amu; (xxi) 1.0000-
CA 02490509 2004-12-17
. ,
-8--
1.0001 amu; (xxii) 1.0001-1.0002 amu; (xxiii) 1.0002-
1.0003 amu; (xxiv) 1.0003-1.0004 amu; (xxv) 1.0004-1.0005
amu; (xxvi) 1.0005-1.0006 amu; (xxvii) 1.0006-1.0007 amu;
(xxviii) 1.0007-1.0008 amu; (xxix) 1.0008-1.0009 amu;
(xxx) 1.0009-1.0010 amu; (xxxi) 0.5 amu; (xxxii) 1.0 amu;
and (xxxiii) 1.0005 amu. The unit amu stands for atomic
mass units (Daltons). A periodicity in the range 0.9990-
1.0010 amu may be observed for singly charged ions and a
periodicity in the range of 0.4995-0.5005 amu may be
observed for doubly charged ions.
One or more of the plurality of regions or portions
of the mass spectral data or the mass spectrum preferably
have a width selected from the group consisting of: (i)
0-0.01 amu; (ii) 0.01-0.02 amu; (iii) 0.02-0.03 amu; (iv)
0.03-0.04 amu; (v) 0.04-0.05 amu; (vi) 0.05-0.06 amu;
(vii) 0.06-0.07 amu; (viii) 0.07-0.08 amu; (ix) 0.08-0.09
amu; (x) 0.09-0.10 amu; (xi) 0.10-0.11 amu; (xii) 0.11-
0.12 amu; (xiii) 0.12-0.13 amu; (xiv) 0.13-0.14 amu; (xv)
0.14-0.15 amu; (xvi) 0.15-0.16 amu; (xvii) 0.16-0.17 amu;
(xviii) 0.17-0.18 amu; (xix) 0.18-0.19 amu; (xx) 0.19-
0.20 amu; and (xxi) > 0.20 amu.
An overall mass window is preferably applied to the
mass spectral data or the mass spectrum. The overall
mass window preferably comprises m nominal mass windows,
wherein m is preferably an integer. According to an
embodiment m may be an even number such that, for
example, m is selected from the group consisting of: (i)
2; (ii) 4; (iii) 6; (iv) 8; (v) 10; (vi) 12; (vii) 14;
(viii) 16; (ix) 18; (x) 20; (xi) 22; (xii) 24; (xiii) 26;
(xiv) 28; (xv) 30; (xvi) 32; (xvii) 34; (xviii) 36; (xix)
38; (xx) 40; (xxi) 42; (xxii) 44; (xxiii) 46; (xxiv) 48;
(xxv) 50; and (xxvi) 52. According to an alternative
and slightly more preferred embodiment, m is preferably
CA 02490509 2004-12-17
. .
- 9 -
an odd number. For example, m may be selected from the
group consisting of: (i) 1; (ii) 3; (iii) 5; (iv) 7; (v)
9; (vi) 11; (vii) 13; (viii) 15; (ix) 17; (x) 19; (xi)
21; (xii) 23; (xiii) 25; (xiv) 27; (xv) 29; (xvi) 31;
(xvii) 33; (xviii) 35; (xix) 37; (xx) 39; (xxi) 41;
(xxii) 43; (xxiii) 45; (xxiv) 47; (xxv) 49; and (xxvi)
51.
According to a less preferred embodiment m may
comprise a fraction.
The nominal mass windows preferably comprise a
substantially contiguous region or portion of the whole
mass spectral data or the mass spectrum. The nominal
mass windows may, less preferably, comprise discrete or
non-contiguous regions or portions of the mass spectral
data or the mass spectrum. One or more of the nominal
mass windows preferably have a width selected from the
group consisting of: (i) 0-0.1 amu; (ii) 0.1-0.2 amu;
(iii) 0.2-0.3 amu; (iv) 0.3-0.4 amu; (v) 0.4-0.5 amu;
(vi) 0.5-0.6 amu; (vii) 0.6-0.7 amu; (viii) 0.7-0.8 amu;
(ix) 0.8-0.9 amu; (x) 0.9-1.0 amu; (xi) 1.0-1.1 amu;
(xii) 1.1-1.2 amu; (xiii) 1.2-1.3 amu; (xiv) 1.3-1.4 amu;
(xv) 1.4-1.5 amu; (xvi) 1.5-1.6 amu; (xvii) 1.6-1.7 amu;
(xviii) 1.7-1.8 amu; (xix) 1.8-1.9 amu; (xx) 1.9-2.0 amu;
and (xxi) > 2 amu.
The nominal mass windows may each have a width
selected from the group consisting of: (i) 0.4995-0.4996
amu; (ii) 0.4996-0.4997 amu; (iii) 0.4997-0.4998 amu;
(iv) 0.4998-0.4999 amu; (v) 0.4999-0.5000 amu; (vi)
0.5000-0.5001 amu; (vii) 0.5001-0.5002 amu; (viii)
0.5002-0.5003 amu; (ix) 0.5003-0.5004 amu; (x) 0.5004-
0.5005 amu; (xi) 0.9990-0.9991 amu; (xii) 0.9991-0.9992
amu; (xiii) 0.9992-0.9993 amu; (xiv) 0.9993-0.9994 amu;
(xv) 0.9994-0.9995 amu; (xvi) 0.9995-0.9996 amu; (xvii)
CA 02490509 2004-12-17
= , =
- 10 -
0.9996-0.9997 amu; (xviii) 0.9997-0.9998 amu; (xix)
0.9998-0.9999 amu; (xx) 0.9999-1.0000 amu; (xxi) 1.0000-
1.0001 amu; (xxii) 1.0001-1.0002 amu; (xxiii) 1.0002-
1.0003 amu; (xxiv) 1.0003-1.0004 amu; (xxv) 1.0004-1.0005
amu; (xxvi) 1.0005-1.0006 amu; (xxvii) 1.0006-1.0007 amu;
(xxviii) 1.0007-1.0008 amu; (xxix) 1.0008-1.0009 amu;
(xxx) 1.0009-1.0010 amu; (xxxi) 0.5 amu; (xxxii) 1.0 amu;
and (xxxiii) 1.0005 amu.
Some or all of the nominal mass windows are
preferably each divided into y channels, wherein y is
preferably selected from the group consisting of: (i) 1;
(ii) 2; (iii) 3; (iv) 4; (v) 5; (vi) 6; (vii) 7; (viii)
8; (ix) 9; (x) 10; (xi) 11; (xii) 12; (xiii) 13; (xiv)
14; (xv) 15; (xvi) 16; (xvii) 17; (xviii) 18; (xix) 19;
(xx) 20; (xxi) 21; (xxii) 22; (xxiii) 23; (xxiv) 24;
(xxv) 25; (xxvi) 26; (xxvii) 27; (xxviii) 28; (xxix) 29;
(xxx) 30; (xxxi) 31; (xxxii) 32; (xxxiii) 33; (xxxiv) 34;
(xxxv) 35; (xxxvi) 36; (xxxvii) 37; (xxxviii) 38; (xxxix)
39; (xl) 40; (xli) 41; (xlii) 42; (xliii) 43; (xliv) 44;
(xlv) 45; (xlvi) 46; (xlvii) 47; (xlviii) 48; (xlix) 49;
(1) 50; and (1i) > 50.
The step of determining an intensity distribution
from a plurality of different regions or portions of mass
spectral data or a mass spectrum preferably comprises
determining the frequency of the various intensities of
the mass spectral data or the mass spectrum in one or
more of the nth channels of one or more of the nominal
mass windows. Preferably, n ranges from 1 to y.
The step of estimating a background intensity for
one or more regions or portions of the mass spectral data
set or mass spectrum from the intensity distribution
preferably comprises determining an x% intensity quantile
from the intensity distribution.
CA 02490509 2004-12-17
,
- 11 -
Preferably, x is selected from the group consisting
of: (i) 0-5; (ii) 5-10; (iii) 10-15; (iv) 15-20; (v) 20-
25; (vi) 25-30; (vii) 30-35; (viii) 35-40; (ix) 40-45;
(x) 45-50; (xi) 50-55; (xii) 55-60; (xiii) 60-65; (xiv)
65-70; (xv) 70-75; (xvi) 75-80; (xvii) 80-85; (xix) 85-
90; (xx) 90-95; and (xxi) 95-100.
The estimated background intensity preferably
comprises the x% intensity quantile or a factor thereof.
The step of adjusting the intensity of one or more
regions or portions of the mass spectral data set or mass
spectrum in order to remove or reduce the effects of the
estimated background intensity preferably comprises
subtracting the estimated background intensity or a
fraction thereof from the one or more regions or portions
of the mass spectral data or mass spectrum. If the
intensity of one or more regions or portions of the mass
spectral data set or mass spectrum has a negative value
or values after substraction of the estimated background
intensity or a fraction thereof, then the intensity of
the one or more regions or portions of the mass spectral
data set or mass spectrum is adjusted or set to zero or
near zero.
The estimated background intensity or a fraction
thereof is preferably subtracted from z% of the mass
spectral data set or the mass spectrum, wherein z is
preferably selected from the group consisting of: (i) 0-
10; (ii) 10-20; (iii) 20-30; (iv) 30-40; (v) 40-50; (vi)
50-60; (vii) 60-70; (viii) 70-80; (ix) 80-90; and (x) 90-
100. The estimated background intensity or a fraction
thereof is preferably subtracted from the one or more
regions or portions of the mass spectral data or the mass
spectrum.
CA 02490509 2004-12-17
. .
- 12 -
The overall mass window is preferably advanced (or
less preferably withdrawn or retreated) one or more
times. For example, the overall mass window may be
advanced or withdrawn at least 1-10, 10-50, 50-100, 100-
150, 150-200, 200-250, 250-300, 300-350, 350-400, 400-
450, 450-500, 500-600, 600-700, 700-800, 800-900, 900-
1000, 100-1250, 1250-1500, 1500-1750, 1750-2000, 2000-
2250, 2250-2500, 2500-2750, 2750-3000 or in excess of
3000 times. According to the preferred embodiment the
overall mass window may be advanced or retreated in steps
of 0.5, 1.0 or 1.0005 atomic mass units (Daltons) or some
other amount each time. It is contemplated the overall
mass window could be advanced or retreated with an
increment selected from the group consisting of: (i)
0.4995-0.4996 amu; (ii) 0.4996-0.4997 amu; (iii) 0.4997-
0.4998 amu; (iv) 0.4998-0.4999 amu; (v) 0.4999-0.5000
amu; (vi) 0.5000-0.5001 amu; (vii) 0.5001-0.5002 amu;
(viii) 0.5002-0.5003 amu; (ix) 0.5003-0.5004 amu; (x)
0.5004-0.5005 amu; (xi) 0.9990-0.9991 amu; (xii) 0.9991-
0.9992 amu; (xiii) 0.9992-0.9993 amu; (xiv) 0.9993-0.9994
amu; (xv) 0.9994-0.9995 amu; (xvi) 0.9995-0.9996 amu;
(xvii) 0.9996-0.9997 amu; (xviii) 0.9997-0.9998 amu;
(xix) 0.9998-0.9999 amu; (xx) 0.9999-1.0000 amu; (xxi)
1.0000-1.0001 amu; (xxii) 1.0001-1.0002 amu; (xxiii)
1.0002-1.0003 amu; (xxiv) 1.0003-1.0004 amu; (xxv)
1.0004-1.0005 amu; (xxvi) 1.0005-1.0006 amu; (xxvii)
1.0006-1.0007 amu; (xxviii) 1.0007-1.0008 amu; (xxix)
1.0008-1.0009 amu; (xxx) 1.0009-1.0010 amu; (xxxl) 0.5
amu; (xxxii) 1.0 amu; and (xxxiii) 1.0005 amu. According
to another embodiment it is contemplated that the overall
mass window could be advanced, withdrawn or translated
(preferably repeatedly) with an increment preferably
selected from the group consisting of: (i) 0-0.1 amu;
CA 02490509 2004-12-17
- 13 -
(ii) 0.1-0.2 amu; (iii) 0.2-0.3 amu; (iv) 0.3-0.4 amu;
(v) 0.4-0.5 amu; (vi) 0.5-0.6 amu; (vii) 0.6-0.7 amu;
(viii) 0.7-0.8 amu; (ix) 0.8-0.9 amu; (x) 0.9-1.0 amu;
(xi) 1.0-1.1 amu; (xii) 1.1-1.2 amu; (xiii) 1.2-1.3 amu;
(xiv) 1.3-1.4 amu; (xv) 1.4-1.5 amu; (xvi) 1.5-1.6 amu;
(xvii) 1.6-1.7 amu; (xviii) 1.7-1.8 amu; (xix) 1.8-1.9
amu; (xx) 1.9-2.0 amu; and (xxi) > 2 amu. According to
other embodiments the overall mass window may be advanced
or withdrawn in regular, non-regular or random steps or
increments.
According to another aspect of the present invention
there is provided a mass spectrometer comprising:
means which determines, in use, an intensity
distribution from a plurality of regions or portions of a
mass spectral data set or mass spectrum;
means which estimates, in use, a background
intensity for one or more regions or portions of the mass
spectral data set or mass spectrum from the intensity
distribution; and
means which adjusts, in use, the intensity of one or
more regions or portions of the mass spectral data set or
mass spectrum in order to remove or reduce the effects of
the estimated background intensity.
The mass spectrometer preferably further comprises
an ion source selected from the group consisting of: (i)
an Electrospray ("ESI") ion source; (ii) an Atmospheric
Pressure Chemical Ionisation ("APCI") ion source; (iii)
an Atmospheric Pressure Photo Ionisation ("APPI") ion
source; (iv) a Laser Desorption Ionisation ("LDI") ion
source; (v) an Inductively Coupled Plasma ("ICP") ion
source; (vi) an Electron Impact ("El") ion source; (vii)
a Chemical Ionisation ("CI") ion source; (viii) a Field
Ionisation ("PT") ion source; (ix) a Fast Atom
CA 02490509 2004-12-17
,
- 14 -
Bombardment ("FAB") ion source; (x) a Liquid Secondary
Ion Mass Spectrometry ("LSIMS") ion source; (xi) an
Atmospheric Pressure Ionisation ("API") ion source; (xii)
a Field Desorption ("FD") ion source; (xiii) a Matrix
Assisted Laser Desorption Ionisation ("MALDI") ion
source; (xiv) a Desorption/Ionisation on Silicon ("DIOS")
ion source; and (xv) a Desorption Electrospray Ionisation
("DESI") ion source.
The ion source may comprise either a continuous ion
source or a pulsed ion source.
The mass spectrometer preferably further comprises a
mass analyser arranged preferably selected from the group
consisting of: (i) an orthogonal acceleration Time of
Flight mass analyser; (ii) an axial acceleration Time of
Flight mass analyser; (iii) a quadrupole mass analyser;
(iv) a Penning mass analyser; (v) a Fourier Transform Ion
Cyclotron Resonance ("FTICR") mass analyser; (vi) a 2D or
linear quadrupole ion trap; (vii) a Paul or 3D quadrupole
ion trap; and (viii) a magnetic sector mass analyser.
The preferred embodiment relates to an adaptive
background subtraction method which reduces the effects
of periodic chemical background noise in mass spectra.
The preferred method examines the intensity
distribution in a local area of a mass spectrum and
estimates that part of the signal due to background noise
by statistical analysis. Further areas of the mass
spectrum are then preferably analysed and the process is
preferably repeated. According to the preferred
embodiment, the estimated background noise in a
particular portion or region of a mass spectrum is
subtracted from the raw or experimentally obtained mass
spectral data to produce a processed mass spectrum which
exhibits significantly reduced background noise. The
CA 02490509 2004-12-17
- 15 -
preferred embodiment is particularly effective in
suppressing background noise having a periodic nature and
also background noise which varies with mass to charge
ratio.
According to the preferred embodiment the intensity
of mass spectral data within a channel of a central
nominal mass window is modified by subtracting an
intensity value from the mass spectral data within the
particular channel. The intensity value which is
subtracted is preferably an intensity quantile (e.g. 45%
or 50%) of the recorded intensities of mass spectral data
within corresponding channels of a plurality of adjacent
or neighbouring nominal mass windows. The preferred
intensity quantile is preferably 45% or 50%, but
according to other embodiments the intensity quantile may
be in the range 10-90%.
The preferred method is particularly suitable for
reducing the effect of background signals which have
periodic intensity variations. The preferred embodiment
is also effective in reducing the effect of unwanted
background noise when the background noise exhibits a
slow continuous variation in intensity relative to the
intensity variation associated with an analyte signal.
The preferred method also enables automated background
subtraction to be performed and enables mass spectra to
be produced which have a significantly improved signal to
noise ratio.
According to an embodiment of the present invention,
a mass spectrum may be divided up into multiple nominal
mass windows preferably centered on multiples of, for
example, 1.0005 atomic mass units (Daltons). An overall
mass window size is preferably chosen which preferably
comprises an odd integer number of nominal mass windows.
CA 02490509 2004-12-17
. ,
- 16 -
The overall mass window size is preferably relatively
large compared to a typical isotope cluster and yet is
also preferably relatively small compared to the low
frequency noise wavelength. According to a particularly
preferred embodiment an overall mass window comprising 21
nominal mass windows may be used wherein the overall mass
window has a width of 21.0105 Da. Each nominal mass
window is preferably 1.0005 Da wide.
Background noise is preferably estimated and then
subtracted from the mass spectral data in one nominal
mass window. Each nominal mass window is preferably
divided into y discrete channels. The width of each
discrete channel y is preferably relatively small
compared to the width of noise peaks and yet is also
preferably relatively large compared to the spacing of
mass spectral data. According to the preferred
embodiment each nominal mass window may be divided up
into 10-20 channels.
The data in the various nominal mass windows which
form the overall mass window can be considered as being
collapsed into y discrete channels per nominal mass
window. An intensity quantile Q of the data across all
the same corresponding channels (i.e. across all the
first, second or nth channels of the nominal mass
windows) is preferably determined at a fraction x%. The
intensity value Q is preferably such that x% of the data
in the respective nth channels of the various nominal
mass windows lies below the intensity value Q. The
intensity quantile is preferably chosen so as to reject
predominantly signal whilst accepting predominantly
noise. The intensity quantile Q is therefore taken to be
a representation of the noise in the corresponding
channel of the central nominal mass window. This
CA 02490509 2004-12-17
. .
- 17 -
background noise is then preferably subtracted from the
input or raw mass spectral data relating to the
corresponding channel of the central nominal mass window.
This process is then repeated for the other channels of
the central nominal mass window. The overall mass window
is then preferably advanced e.g. approximately 1 atomic
mass unit, and the process is preferably repeated,
preferably multiple times.
The calculated background distribution may, for
example, contain data distributed across 20 channels per
mass unit whereas the raw mass spectral data may contain
many more data points per mass unit. In the case of time
of flight data the number of data points per mass unit
will vary. The intensity of the estimated background
noise which is to be subtracted at a given data point in
the original mass spectral data may be calculated by
interpolation between the 20 data points which form the
estimated background distribution across a nominal mass
window having a width of approximately 1 atomic mass
unit.
The preferred method has the particular advantage
over the known frequency domain filtering method in that
the preferred method avoids creating artefacts or extra
noise spikes in the processed mass spectral data. Such
artefacts or extra noise spikes can be a particular
problem when the known frequency domain filtering
approach is used.
Various embodiments of the present invention will
now be described, by way of example only, and with
reference to the accompanying drawings in which:
Fig. lA shows a portion of a mass spectrum
exhibiting repetitive chemical noise with an overall mass
window comprising nine nominal mass windows each divided
CA 02490509 2004-12-17
- 18 -
into ten channels superimposed upon the portion of the
mass spectrum, and Fig. 13 shows the intensity
distribution of all the intensity data taken from all the
first channels of the nine nominal mass windows shown in
Fig. 1A;
Fig. 2 shows in greater detail the central nominal
mass window M5 and the immediately adjacent nominal mass
windows M4,M6 as shown in Fig. 1A together with the
calculated background noise for most of the central
nominal mass window M5, and the inset shows in greater
detail the analyte mass peak having a mass to charge
ratio of approximately 647.6 after removal of background
noise according to the preferred embodiment;
Fig. 3A shows a portion of a mass spectrum
exhibiting periodic background noise and Fig. 3B shows
the same portion of the mass spectrum after removal of
the periodic background noise according to the preferred
embodiment;
Fig. 4A shows in greater detail a portion of the
mass spectrum shown in Fig. 3A, Fig. 4B shows the same
portion of the mass spectrum after removal of the
periodic background noise according to the preferred
embodiment and Fig. 40 shows the estimated periodic
background noise which was subtracted from the
unprocessed mass spectrum shown in Fig. 4A;
Fig. 5A shows a portion of a mass spectrum
exhibiting periodic background noise and Fig. 5B shows
the same portion of the mass spectrum after removal of
the periodic background noise according to the preferred
embodiment;
Fig. 6A shows a portion of a mass spectrum
exhibiting slowly continuous and periodic background
noise and Fig. 6B shows the same portion of the mass
CA 02490509 2004-12-17
- 19 -
spectrum after removal of the slowly continuous and
periodic background noise according to the preferred
embodiment; and
Fig. 7A shows in greater detail a portion of the
mass spectrum shown in Fig. 6A, and Fig. 7B shows the
same portion of the mass spectrum after removal of the
slowly continuous and periodic background noise according
to the preferred embodiment.
An embodiment of the present invention will now be
described with reference to Figs. 1A, 1B and 2. However,
the embodiment shown and described with reference to
Figs. 1A, 1B and 2 has been simplified for ease of
illustration. According to a particularly preferred
embodiment, an overall mass window having a width of
21.0105 atomic mass units (Daltons) and comprising 21
nominal mass windows each 1.0005 atomic mass units
(Daltons) wide is applied to a mass spectrum. Each
nominal mass window is preferably divided into 20
discrete channels. However, for ease of illustration,
the embodiment shown and described with reference to
Figs. 1A, 1B and 2 relates to using a smaller overall
mass window which is only 9 atomic mass units wide and
which comprises only 9 nominal mass windows each having a
width of precisely 1 atomic mass unit (Dalton). Each
nominal mass window is shown divided into 10 discrete
channels, again for ease of illustration.
Fig. lA shows a portion of a mass spectrum across
the mass to charge ratio range 643-652 which exhibits
repetitive or periodic chemical background noise. The
mass spectrum was obtained using an Electrospray ion
source and a Time of Flight mass analyser. An overall
mass window having a width of 9 atomic mass units is
shown superimposed upon the portion of the mass spectrum.
CA 02490509 2004-12-17
,
- 20 -
The overall mass window comprises nine nominal mass
windows Ml-M9. The nominal mass windows Ml-M9 are
centred around a central nominal mass window M5 which
corresponds to the mass to charge ratio range 647-648.
Each of the nine nominal mass windows Ml-M9 is shown sub-
divided into ten equal width discrete channels a-j. In
the particular example shown in Fig. 1A, each discrete
channel covers or includes approximately 15 mass
intensity pairs or data points. The number of mass
intensity pairs or data points per channel depends upon
the digitisation rate of the acquisition electronics and
the time of flight of the ions analysed.
According to the preferred embodiment, a background
signal is estimated and is then subtracted from the raw
intensity data corresponding to the first channel M5a of
the central nominal mass window M5. The estimated
background signal for the first channel M5a of the
central nominal mass window M5 is calculated by first
determining the intensity distribution of the Intensity
data in or across all the first channels Mla-M9a of all
nine nominal mass windows Ml-M9 which form the overall
mass window. The first channels M1a-M9a of each of the
nine nominal mass windows Ml-M9 are shown as shaded areas
in Fig 1A.
Fig. 1B shows the resulting intensity distribution
which corresponds with or represents the intensity data
taken from all of the first channels Mla-M9a of the nine
nominal mass windows Ml-M9 shown in Fig. 1A. In total,
the first channels Mla-M9a comprise, cover or include
approximately 134 data points or distinct intensity
measurements. A dotted line indicates the 50% intensity
quantile for the intensity distribution shown. In the
particular example shown in Fig. 1B, the 50% intensity
CA 02490509 2004-12-17
- 21 -
quantile is 19. A 50% intensity quantile represents an
intensity value wherein 50% of the recorded intensities
lie below the 50% quantile and 50% of the recorded
intensities lie above the 50% quantile. Accordingly, 50%
of the intensity data in all the first channels Mla-M9a
of the nine nominal mass windows M1-M9 shown in Fig. 1A
has an intensity value 19 units and 50% of the
intensity data in all the first channels Mla-M9a of the
nine nominal mass windows M1-M9 has an intensity value
19 units. Other embodiments are contemplated wherein
intensity quantiles other than 50% are used. For
example, an intensity quantile of 45% may be used wherein
45% of the intensity data has an intensity less than or
equal to the 45% intensity quantile. In the particular
example shown in Fig. 18 the 45% intensity quantile would
have a value of 18.
In the particular example shown and described with
reference to Figs. 1A, 113 and 2, the 50% intensity
quantile value of 19 is deemed as being representative of
the average intensity of the background signal in the
first channel M5a of the central nominal mass window M5.
The 50% intensity quantile value of 19 is therefore
preferably subtracted from the intensity values of all
the raw intensity data which make up or fall within the
first channel M5a of the central nominal mass window M5.
According to an embodiment if the intensity or the
intensity values are negative or take a negative value
after subtraction of the intensity quantile value, then
preferably the intensity or intensity values are set or
adjusted to zero, or less preferably to a value close to
zero.
Having determined the predicted intensity of the
background signal for the first channel M5a of the
CA 02490509 2004-12-17
,
- 22 -
central nominal mass window M5, this procedure is then
preferably repeated for the second channel M5b of the
central nominal mass window M5. In a similar manner, the
50% intensity quantile for the intensity distribution
relating to all the intensity data taken from all the
second channels Mlb-M9b of the nine nominal mass windows
Ml-M9 is preferably determined. This intensity value is
then preferably deemed as being representative of the
average intensity of the background signal in the second
channel M5b of the central nominal mass window M5. This
new 50% intensity quantile value is then preferably
subtracted from the intensity values of all the raw
intensity data which make up or fall within the second
channel M5b of the central nominal mass window M5.
This process is then again preferably repeated in a
similar manner for the third, fourth, fifth, sixth,
seventh, eighth and ninth channels M5c-M5j of the central
nominal mass window M5. As a result, the background
noise is preferably estimated across the whole of the
width of the central nominal mass window M5 and the
estimated background noise is then preferably duly
subtracted from the raw intensity data in all ten
channels M5a-j of the central nominal mass window M5.
According to an embodiment, if the intensity data after
subtraction of the estimated background noise takes or
has a negative value then preferably the intensity data
is adjusted or set to zero, or less preferably near zero.
After having calculated and then preferably removed
the estimated background noise from the raw intensity
data relating to the central nominal mass window M5 which
relates to ions having mass to charge ratios in the range
647-648, the overall mass window is then preferably
advanced approximately one mass unit further on (or less
CA 02490509 2004-12-17
- 23 -
preferably retracted approximately one mass unit
backwards) so that the overall mass window is now
preferably centred on the mass to charge ratio range of
648-649. The procedure as described above in relation to
determining the background noise across the previous
central nominal mass window and mass to charge ratio
range 647-648 is then preferably repeated in order now to
estimate the background noise across the new central
nominal mass window and mass to charge ratio range 648-
649. The estimated background noise is then preferably
removed from the raw intensity data corresponding to the
new central nominal mass window which covers the mass to
charge ratio range 648-649. The overall mass window is
then preferably advanced approximately one mass unit
further on (or less preferably retracted approximately
one mass unit backwards) and the process of estimating
the background noise and subtracting the estimated
background noise from the new central nominal mass window
is then preferably repeated.
The process of determining the background noise and
subtracting the background noise from the central nominal
mass window and then advancing, withdrawing, translating
or otherwise moving the overall mass window is then
preferably repeated until background noise has been
removed from the region, portion or whole of the mass
spectrum of interest. According to an embodiment the
overall mass window may be advanced, withdrawn or
translated at least 10, 50, 100, 200, 300, 400, 500, 600,
700, 800, 900, 1000, 1500, 2000, 2500, 3000, 3500, 4000,
4500 or 5000 times. The width of the overall mass window
preferably stays the same, but according to other
embodiments the width of the overall mass window may
CA 02490509 2004-12-17
- 24 -
increase, decrease or otherwise vary in a stepped,
linear, random or other manner.
Fig. 2 shows the central nominal mass window M5 and
the immediately neighbouring nominal mass windows M4,M6
as shown in Fig. 1A in greater detail. The calculated
background noise for most of the central nominal mass
window M5 is shown superimposed upon the original or raw
intensity or mass spectral data. The inset shows in
greater detail a portion of the central nominal mass
window M5 after the intensity data has been processed to
subtract the calculated or estimated background noise
therefrom. An analyte mass peak having a mass to charge
ratio of approximately 647.6 can be seen more clearly and
the signal to noise ratio of the processed mass spectrum
has been significantly improved.
Fig. 3A shows a portion of a mass spectrum
exhibiting periodic background noise having a periodicity
of approximately 1 atomic mass unit. The mass spectrum
was obtained using an Electrospray ion source and a Time
of Flight mass analyser. Although intense analyte peaks
can be identified, it is difficult to discern
comparatively weaker analyte peaks from amongst the
periodic background noise. Fig. 3B shows the same
portion of the mass spectrum after the periodic
background noise has been estimated and subtracted from
the intensity data according to the preferred embodiment.
In this particular example the overall mass window
applied to the mass spectral data comprised 21 nominal
mass windows each having a width of 1.0005 atomic mass
units (Daltons). Each nominal mass window was divided
into 20 channels and a 45% intensity quantile was used to
discriminate between signal and background noise.
CA 02490509 2004-12-17
. ,
- 25 -
Fig. 4A shows in greater detail a portion of the
mass spectrum shown in Fig. 3A over the mass to charge
ratio range 934-956. As can be seen, the intensity of
some of the analyte mass peaks are not significantly
greater than the intensity of some of the peaks due to
periodic background noise. It is to be noted, for
example, that peak recognition software has suggested
that peaks observed having mass to charge ratios of
944.7, 953.7 and 955.7 are analyte peaks. However, in
fact, these peaks are believed to be peaks due to
background noise. Fig. 4B shows a corresponding mass
spectrum after the periodic background noise has been
estimated and subtracted from the intensity data shown in
Fig. 4A. Analyte peaks having mass to charge ratios of
937.5, 938.5, 947.7 and 948.7 are now more clearly
identifiable as being analyte peaks. Furthermore, the
peaks observed in Fig. 4A which were determined as having
mass to charge ratios of 944.7, 953.7 and 955.7 have now
been substantially suppressed as background noise. Fig.
40 shows the intensity of the periodic background noise
as calculated or estimated according to the preferred
embodiment for the intensity data shown in Fig. 4A. The
background noise shown in Fig. 40 was removed or
otherwise subtracted from the raw mass spectral data
shown in Fig. 4A to produce the improved processed mass
spectrum shown in Fig. 4B.
Fig. 5A shows another portion of a mass spectrum
exhibiting periodic background noise having a periodicity
of approximately 1 atomic mass unit. The mass spectrum
was obtained using an Electrospray ion source and a Time
of Flight mass analyser. Fig. 5B shows the resultant
mass spectrum after the periodic background noise had
been calculated or estimated and subtracted from the
CA 02490509 2004-12-17
- 26 -
intensity data according to the preferred embodiment. An
overall mass window comprising 21 nominal mass windows
each 1.0005 atomic mass units (Daltons) wide was applied
to the mass spectral data. Each nominal mass window was
divided into 20 channels and a 45% intensity quantile was
used to discriminate between signal and background. As
can been seen from Fig. 5B, the periodic background noise
has been strongly suppressed and the signal to noise
ratio has been significantly enhanced or improved.
Fig. 6A shows a mass spectrum of the tryptic digest
products of a protein ionised by a Matrix Assisted Laser
Desorption Ionisation ion source and mass analysed by an
axial Time of Flight mass analyser. The mass spectrum
exhibits both slowly varying background noise and also
periodic background noise (as can be seen more clearly in
Fig. 7A). Fig. 6B shows the resultant mass spectrum
after the slowly varying background noise and also the
periodic background noise had been calculated or
estimated and subtracted from the intensity data
according to the preferred embodiment. An overall mass
window comprising 21 nominal mass windows each 1.0005
atomic mass units (Daltons) wide was applied to the mass
spectral data. Each nominal mass window was divided into
20 channels and a 45% intensity quantile was used to
discriminate between signal and background.
Fig. 7A shows in greater detail a portion of the
mass spectrum shown in Fig. 6A over the mass to charge
ratio range 1754-1789. Fig. 7B shows the resultant mass
spectrum after the slowly varying background noise and
the periodic background noise had been estimated and
subtracted from the intensity data according to the
preferred embodiment. As can be seen from Fig. 78, the
CA 02490509 2004-12-17
. .
- 27 -
background noise has been strongly suppressed and the
signal to noise ratio has been significantly improved.
The preferred embodiment is particularly effective
in reducing the undesired effects of chemical background
noise having a periodicity of approximately 1 atomic mass
unit as is commonly observed in mass spectra at mass to
charge ratios less than 2000. Other embodiments are also
contemplated wherein different width nominal mass windows
and/or a different number of channels per nominal mass
window may be used particularly when the background noise
in a mass spectrum exhibits a periodicity other than
approximately 1 atomic mass unit and/or when the
background noise exhibits a more complex nature.
Embodiments are also contemplated which are intended
to filter out background noise which has two or more
characteristic repeat periods. According to such
embodiments, the format of the overall mass window
applied to a mass spectrum may be modified or may vary so
that the mass spectral data in, for example, only odd,
even or every nth numbered nominal mass windows are
sampled when determining background noise. Embodiments
are also contemplated wherein each nominal mass window
may have a width of, for example, 0.5 atomic mass units
or some other value other than 1.
According to another embodiment in order to speed up
the processing time, the intensities within one or more
channels of a nominal mass window may be averaged prior
to calculating the desired intensity quantile.
As mentioned above, it is possible that the
intensity data after substraction of an intensity
quantile could be determined as having a negative value.
In such circumstances the intensity data is then
preferably set or adjusted to zero or to near zero.
CA 02490509 2004-12-17
. ,
- 28 -
Although the present invention has been described
with reference to preferred embodiments, it will be
understood by those skilled in the art that various
changes in form and detail may be made without departing
from the scope of the invention as set forth in the
accompanying claims.
