See github.com/pwilmart/ASM... for my thoughts. Typically (for DDA), some parameters affect scoring and those should be optimized for each search tool. Other parameters affect search space size (FASTA file, parent tolerance, trypsin cleavage, PTMs) and search space should be the same (or similar).
#proteomics_101 When comparing results from different PSM assignment algorithms, is it best to use parameters optimized for each individual algorithm or a single set of generic parameters for all of them? What are the best ways to compare the results, both overall and in detail?
Unstated: about half of this money is Americans buying Canadian companies.
If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
pluralistic.net/2026/04/20/p...
2/
A NanoBeast Z80 computer driving a 14 segment LCD display with a rather shiny backlight.
I've just spent four hours tracking down a software bug...
...that turned out to be a dry solder joint.
On the one hand, I'm quite annoyed. On the other hand I'm very happy - it works!!! It's taken a while to get here, but I've got a properly working LCD screen. (Spot the bodges!)
#proteomics_101 If you were reviewing a study that claims to have found patterns of protein concentration in biobanked plasma indicative of a specific cancer, what statistical tests would you use to determine whether any pattern was caused by the cancer and not its treatment?
A line graph of new and competitive renewal grants for fiscal years 2015 through 2026. The curve for 2026 lies well below that for previous fiscal years.
The story includes a funding curve graph showing the delay in making new and competitive renewal grants.
2/8
Lagging? 2026 has flattened at 33% of normal! This is devastating.
I love it when someone responds to a question about the merits of an LLM model by saying that it is "more powerful". It means I don't have to listen to the rest of the interview.
One of my favourite stats tests, although it is under-utilized for some reason.
I would use a Kolmogorov-Smirnov test against a uniform distribution.
#proteomics_101 Human PSD3:p has what appear to be too many clusters of 3 Y/F (blue) residues in its cward pIDRs. What statistical tests would you use to determine if these were a potentially interesting or a simply stochastic phenomenon?
It is unusual to have 2 asteroids get closer than a lunar distance from the earth on successive days.
Americans, since the beginnings of their nation, have been big fans of DIY religious creeds & a significant fraction of the population consider them to be every bit as valid as any established authority's opinion.
#proteomics_101 Human MT-ND4L:p is one of the 13 protein sequences encoded on the mitochondrial chromosome. It does not have any observable tryptic peptides, but peptides from its sequence have been observed. What type of experiment would result in finding these peptides?
#proteomics_101 Provide at least 2 testable hypotheses that could explain why there is such a difference in peptide frequency of observation for yeast's SED4:p between the regions (1-400) & (401-1065)?
I love it when they just make up stuff in the hopes that enough people will believe it ...
nypost.com/2026/04/07/u...
Canadian advocacy groups have been sending me messages to contact my MP to prevent the passage of Bill C22. Please stop doing this. MPs are worse than useless. Instead, tell me the names of the lobbyists who are pushing the bill & their clients so I can apply pressure where it might be effective.
#proteomics_101 Human SPCS1:p has 4 exons & 102 residues, supported by both proteomics & nucleic acid methods. UniProt lists a 169 aa sequence as canonical that has not been observed. What steps would you have to go through to correct this issue? Estimate the likelihood of success at each step.
Do the newer proteomics platforms (Olink, SomaLogic, etc) use the frequently-glitchy UniProt sequence collection or have they adopted a more up-to-date system?
I am over-fond of QR codes. I once put together a conference poster that was multiple QR codes & captions. The codes pointed to a web site with videos & slide decks illustrating the information associated with captions. Unfortunately, I don't think anyone passing by actually pulled out their phone.
#proteomics_101 Identify the signal sequence & mature N-terminus of the protein sequence given below.
The bad ideas just keep on coming. GLP-1 is bad enough: it has many metabolically active buddies that are even worse.
www.cnn.com/2026/04/13/h...
Start at around 16:30 to listen to the former CFO of Google suggesting our country take the Orban-esque step of slapping an exit tax on Canadians who work abroad so as to repay the cost of their education. And listen to all the Liberals clap like fucking seals at this appalling illiberal nonsense.
