missing field

[AUldry]

Dear all,

I wanted to try this interesting piece of software but when I ran it with the exact same config file (bar for the path to the fasta file, which is adjusted to my set up)

ionmesh -c ionmesh_config.toml test.d

I get the following message:

thread 'main' panicked at src/main.rs:97:49:
called Result::unwrap() on an Err value: Error { inner: Error { inner: TomlError { message: "missing field max_mz_expansion_ratio", raw: Some("[denoise_config]\r\nmz_scaling = 0.015\r\nims_scaling = 0.03\r\nms2_min_n = 2\r\nms1_min_n = 3\r\nms1_min_cluster_intensity = 100\r\nms2_min_cluster_intensity = 50\r\n\r\n[tracing_config]\r\nmz_scaling = 0.019999999552965164\r\nrt_scaling = 2.200000047683716\r\nims_scaling = 0.02\r\nmin_n = 2\r\nmin_neighbor_intensity = 200\r\n\r\n[pseudoscan_generation_config]\r\nrt_scaling = 0.7\r\nquad_scaling = 5.0\r\nims_scaling = 0.02\r\nmin_n = 4\r\nmin_neighbor_intensity = 500\r\n\r\n[sage_search_config]\r\nstatic_mods = [[\r\n "C",\r\n 57.0215,\r\n]]\r\nvariable_mods = [[\r\n "M",\r\n [15.9949],\r\n]]\r\nfasta_path = "myfasta.fasta"\r\n\r\n[output_config] # These options can be missing, if missing will not output the files.\r\nout_features_csv = "features.csv"\r\ndebug_traces_csv = "debug_traces.csv"\r\ndebug_scans_json = "debug_scans.json"\r\n"), keys: ["denoise_config"], span: Some(0..151) } } }
note: run with RUST_BACKTRACE=1 environment variable to display a backtrace

Any idea what I do wrong? Many thanks in advance!

Anne-Christine

[jspaezp]

Hello! Indeed It seems like I forgot to update the documentation on the new version of the config file! Let me get back to you on this one later in the day!

Although I believe these are all the fields you might need right now... (I am working on re-parametrizing them and finding good defaults so they don't have to be specified by hand ...)

[denoise_config]
mz_scaling = 0.015
max_mz_expansion_ratio = 1.5
ims_scaling = 0.015
max_ims_expansion_ratio = 5.0
ms2_min_n = 5
ms1_min_n = 10
ms1_min_cluster_intensity = 100
ms2_min_cluster_intensity = 50

[tracing_config]
mz_scaling = 0.015
max_mz_expansion_ratio = 1.5 # This is the one throwing the error .. (I will also make a better error message ...)
rt_scaling = 2.4
max_rt_expansion_ratio = 3.0 # But this one was added at the same time
ims_scaling = 0.02
max_ims_expansion_ratio = 4.0 # And this one ...
min_n = 3
min_neighbor_intensity = 450

[pseudoscan_generation_config]
rt_scaling = 2.4
max_rt_expansion_ratio = 5.0 # ... and this one
quad_scaling = 5.0
max_quad_expansion_ratio = 1.0 # ... 
ims_scaling = 0.015
max_ims_expansion_ratio = 2.0 # ...
min_n = 6
min_neighbor_intensity = 6000

[sage_search_config]
static_mods = [[
    "C",
    57.0215,
]]
variable_mods = [
    ["M", [15.9949]],
]
fasta_path = "UP000005640_9606.fasta"

[AUldry]

Many thanks for your reply!

I think the config file does need the output_config, as otherwise we get a "missing field output_config" error.

The software now runs, but now I have the problem that I need to find a suitable example to test, or a suitable machine to run on, as I seem to eventually go on my local machine into a memory problem (at least I believe that's what causes the "Killed" output). What RAM would you say I need for, say, a 7.3 GB .d sample?

[jspaezp]

Well ... it depends ...

If we break down the process in 'processing' and 'search'.

• On the 'processing' stage the peak consumption should be when all the frames are stored at the same time without denoising ... and I would expect that to be very similar to the size of the file. (which raises a good point ... there might be a more efficient way of doing this ... I could make a 'low memory' mode that loads frames one at a time, which might be slower but could cut memory usage by half).
• On the search stage ... it greatly depends on the search parameters... since the underlying search engine is sage, the peptide database needs to fit in memory. Empirically ... a human database with trypsin and 1 missed cleavage takes ~4gb If i recall correctly, but variable modifications make the memory requirements explode (+30gb for M-ox + STY-phospho if 2 mods are allowed per peptide).

I test everything on a 16 GB mac air ... which so far has worked for 44 min diaPASEF data on human with those mods.
If you send me an email I could send you one of the raw files I use for testing (jspaezp1 in twitter if you dont want to post it here)

best!

[AUldry]

Dear Sebastian,

I would gladly take up your offer of trying your own example. My email is anne-christine.uldry followed by at, foloowed by unibe.ch (hope it's clear!). Thanks!