Superresolution Chromatography

E.L.Kosarev - P.L.Kapitza Institute for Physical Problems, RAS

http://www.kapitza.ras.ru/people/kosarev/home.htm

K.O.Muranov - N.M.Emanuel Institute of Biochemical Physics, RAS

http://kmuranov.euro.ru

Chromatographic analysis

Chromatography: Adsorption/Partition Affinity Ion-exchange Size-exclusion

Peaks’ overlapping problem

Peaks overlapped form the joint peak

Retention time

Sa

mp

le c

on

cen

tra

tion

Imperfects of parametric deconvolution

• Supposition - Peak shape can be described analytically (Gaussian, Lorenz, etc.)

• Parametric deconvolution needs:– peak position

– peak wide

– peak amplitude.

• Unknown peak characteristic causes an error.

For instance, two peaks were identified instead of three

0 500 1000 1500 2000

0.00

0.01

0.02

0.03

0.04

Retention time

Imperfects of parametric

deconvolution

Peak broadeningbroadening and column voidingcolumn voiding

The main factors Diffusion Non-specific interaction System overdampening (irreversible adsorption,

filter contamination, plunger damage, etc.)

Retention time

Ab

sorb

an

ce

Problems:

• the shape of a real chromatography peak could not be approximated exactly with any function

• the parametric technique cannot give reliable results

We suppose:

the problem of overlapping peaks’ separation could be accurately solved

with the use of a nonparametric

method

Nonparametric method

• The peak's shape is determined directly from the separation of an individual compound and can be called the point spread function of a chromatographic column

• If the shapes of these peaks are the same throughout the entire working range of the device, then chromatogram is a distribution convolution with the peak of this shape

• the point spread function includes all factors influenced the separation

Nonparametric method

• The decomposing a complex spectrum into the same components is achieved by solving an integral convolution equation

• For this equation: – the input data is the chromatogram– the convolution operator kernel is the point spread

function of the chromatograph column

Analysis of chromatography separation data with the

RECOVERY software package

Method

• A protein mixture of known composition (bovine serum albumin monomer, dimer, and trimer) was separated by gel filtration for obtaining of heavily overlapping peaks.

• The data was processed with the use of the RECOVERY software package

• The result was compared with the finer separation data

obtained with the use of the HPLC.

BSA chromatography and point spread function determination

A - Elution profile of bovine serum albumin (BSA).

The blue arrow indicates unresolved peak

Red dashed lines mark the time interval when the BSA monomer fraction was collected

B - Elution profile of the collected fraction

Absence of unresolved peak Strongly pronounced a peak

asymmetry

Recovering of the chromatographic separation

data

• A - the source BSA chromatogram;

• B - point-spread function;

• C - recovering result of the chromatographic separation data with the RECOVERY software package.

The RECOVERY result vs. HPLC separation

A - data recovered with nonparametric approach

B - Elution profile of HPLC separation (TSK G2000 SW Spherogel, 10mm X 600 mm, flowrate - 0.5ml/min)

• Both Recovery software package and HPLC found the BSA monomer, dimer and trimer in protein mixture

• Recovery shows the more precision result

Peak width and whole column limit

- under investigation

• Separation of the rat eye lens crystallins (magenta)– 10 kD - 1.5 MD

limit• Peak width of the

standard proteins (blue) Retention time, sec

Pro

tein

con

cent

ratio

n

-200

200

600

1000

1400

1800

2000 4000 6000 8000 10000 12000

Conclusion:

• the proposed method fundamentally improves the quality of the chromatographic separation– RECOVERY software package for the gel filtration data

significantly increased the resolution of this method and exceeded the quality of the separation obtained with the HPLC technique

• new possibilities are achieved through reasonable processing of the measured data with no complication in the instrumentation– cost of the used instrument complex for gel filtration

($1,000) is roughly 15-20 times lower than that for the HPLC setup ($20, 000).