Unlock Proteomics Gold
Unlock
Proteomics Gold
from Data Independent Acquisition (DIA) Mass Spectrometry (MS) Data
Please scroll down
Proteins Determine Phenotype
DIA-MS Does Discovery Proteomics
Yet, ~75-95% of Peptides in DIA-MS Files Are Ignored!
Imagine Unlocking All that Unused "Proteomics Gold" In DIA-MS Data
Many labs have acquired DIA-MS data over the years but have been unable to access ~75-95% of the peptides present in the DIA-MS files. Moreover, this ~75-95% set exclusively contains the peptides with unpredicted sequences and unexpected post translational modifications (PTMs) that are — almost by definition — far more likely to separate study conditions.
What if we could unlock the hidden value from your past, present, and future DIA-MS data?
For nearly 30 years, discovery-MS proteomics has been promising to outperform genomics, and for good reasons. What if we can help your lab start delivering on that promise? That is, what if, once we quantify the set of ~75-95% of peptides in DIA-MS files that were previously unquantified, either your lab or ours can use straightforward AI to create a parsimonious panel that separates study conditions?
How might that impact science, publications, patent submissions, your lab, your career, or the general public's lives?
Many labs have acquired DIA-MS data over the years but have been unable to access ~75-95% of the peptides present in the DIA-MS files. Moreover, this ~75-95% set exclusively contains the peptides with unpredicted sequences and unexpected post translational modifications (PTMs) that are — almost by definition — far more likely to separate study conditions.
What if we could unlock the hidden value from your past, present, and future DIA-MS data?
For nearly 30 years, discovery-MS proteomics has been promising to outperform genomics, and for good reasons. What if we can help your lab start delivering on that promise? That is, what if, once we quantify the set of ~75-95% of peptides in DIA-MS files that were previously unquantified, either your lab or ours can use straightforward AI to create a parsimonious panel that separates study conditions?
How might that impact science, publications, patent submissions, your lab, your career, or the general public's lives?
Peptides containing unpredicted sequences or unexpected PTMs are up to ~2000% more numerous than peptides identified from a FASTA library search space*.
Peptides containing unpredicted sequences or unexpected PTMs are up to ~2000% more numerous than peptides identified from a FASTA library search space*.
These unexpected PTMs or unpredicted sequences are far more likely to create a parsimonious panel that separates your study conditions**.
These unexpected PTMs or unpredicted sequences are far more likely to create a parsimonious panel that separates your study conditions**.
View our preprint or book a web call with our engineers
View our preprint or book a web call with our engineers
* Please see reference to Mann lab study for DDA data (from 2011!) in preprint (top of page 3)
** Please see reference to Alzheimer study in preprint (top of page 3) as well as supplemental note #S5
Our Innovation:
The best way to learn about our novel, patent pending innovation is through an interactive web call. The preprint is a good resource too. Alternatively, for a very quick overview, please see the four figures below.
Global XIC Deconvolution
Our Innovation:
The best way to learn about our novel, patent pending innovation is through a web call. The preprint is a good resource too. Alternatively, for a quick overview, please click on the four figures to your right.
Global XIC Deconvolution
01
PROBLEM
DIA Produces Chimeric XICs
Two co-eluting peptides from a single sample
02
PROPOSED SOLUTION
LC has Natural Variance
Pairs of peptides that co-elute in one subset of samples do not exactly co-elute in another subset of samples
03
MULTIPARTITE MATCHING
Deconvolute *MS2* Fragments Computationally
Match fragment accross samples to create one clean spectra per peptide
04
QUANTIFICATION & AI
Quantify & Create Predictive Panel
Quantify all analytes in MS and use AI to create predictive panel
Our Innovation:
The best way to learn about our novel, patent pending innovation is through a web call. The preprint is a good resource too. Alternatively, for a quick overview, please click on the four figures to your right.
Global XIC Deconvolution
01
PROBLEM
DIA Produces Chimeric XICs
Two co-eluting peptides from a single sample
02
PROPOSED SOLUTION
LC has Natural Variance
Pairs of peptides that co-elute in one subset of samples do not exactly co-elute in another subset of samples
03
MULTIPARTITE MATCHING
Deconvolute *MS2* Fragments Computationally
Match fragment accross samples to create one clean spectra per peptide
04
QUANTIFICATION & AI
Quantify & Create Predictive Panel
Quantify all analytes in MS and use AI to create predictive panel
Ready to find those peptides with unexpected sequences or PTMs that that separate study conditions?
Ready to possibly
publish or patent the results?
The Power Of One
The Power Of One
Although we have observed up to ~2000% more quantified peptides than only identified ones, it matters little whether this increase is ~2000%, ~20000%, or ~20%.
Instead, the key argument is this: unexpected sequences and PTMs are orders-of-magnitude more biologically valuable in separating study conditions (almost by definition) than generic sequences found in a FASTA library search space (with only a few PTMs considered instead of hundreds), yet they are no more / no less likely to be seen by the MS than those generic FASTA sequences (i.e., the MS is unbiased). So, if there is reasonable chance that those biologically valuable but unexpected sequences or PTMs exist in your samples / MS, we at GoldenHaystack Lab can quantify them — and then the resulting follow-up work is relatively straightforward.
And, even if we were to temporarily pretend that there were only a few such unexpected peptide sequences or PTMs in the DIA-MS files — in truth, there are thousands (i.e., they represent the vast majority of analytes in the DIA-MS files, never mind that they are the biologically invaluable ones) — one only needs one or two "good" ones to revolutionize a specific field of study.
Although we have observed up to ~2000% more quantified peptides than only identified ones, it matters little whether this increase is ~2000%, ~20000%, or ~20%.
Instead, the key argument is this: unexpected sequences and PTMs are orders-of-magnitude more biologically valuable in separating study conditions (almost by definition) than generic sequences found in a FASTA library search space (with only a few PTMs considered instead of hundreds), yet they are no more / no less likely to be seen by the MS than those generic FASTA sequences (i.e., the MS is unbiased). So, if there is reasonable chance that those biologically valuable but unexpected sequences or PTMs exist in your samples / MS, we at GoldenHaystack Lab can quantify them — and then the resulting follow-up work is relatively straightforward.
And, even if we were to temporarily pretend that there were only a few such unexpected peptide sequences or PTMs in the DIA-MS files — in truth, there are thousands (i.e., they represent the vast majority of analytes in the DIA-MS files, never mind that they are the biologically invaluable ones) — one only needs one or two "good" ones to revolutionize a specific field of study.
We Unlock Potentially Astronomical Value From
Past, Present, and Future DIA-MS Data
We Unlock Potentially Astronomical Value From
Past, Present, and Future DIA-MS Data
We quantify the set of peptides that exclusively contain the unpredicted sequences and unexpected PTMs (the ~75- 95% of peptides in DIA-MS, an up-to ~2000% increase). These peptides are — almost by definition — far more likely to form a parsimonious panel that separates study conditions*.
We quantify the set of peptides that exclusively contain the unpredicted sequences and unexpected PTMs (the ~75- 95% of peptides in DIA-MS, an up-to ~2000% increase). These peptides are — almost by definition — far more likely to form a parsimonious panel that separates study conditions*.
We seek labs interested in (a) moving quickly, (b) thinking big, and (c) collaborating well.
We seek labs interested in (a) moving quickly, (b) thinking big, and (c) collaborating well.
*Please see reference to Alzheimer study on top of page three of preprint as well as supplemental note #S5.
(C) Copyright GoldenHaystack Lab 2025. All Rights Reserved.
(C) Copyright GoldenHaystack Lab 2025. All Rights Reserved.