Search results

I have no idea how they came to this result. They mention it in the text and have a Figure 5 explaining it but I don't see how this came out of their AI model. Perhaps it's based on further correlation analyses within each group?

All datasets and the full code of their model is available on Github: https://github.com/ohlab/BioMapAI/tree/main

This graph (Extended Data Fig. 3) is also interesting as it shows the most useful biomarkers for each of the 12 symptoms outcomes. These are based on SHAP (SHapley Additive exPlanations) values which indicate how much each feature contributed to a specific prediction.

Rather than just predicting ME/CFS versus control, the model could also estimate symptom scores such as subscales of the SF-36. It seems that the immune data was most useful for prediction physical functioning and general health while info on gut microbiome was useful for gastrointestinal...

Supplementary table 3 gives some more info about the predictions. Using all datasets (metabolomics, immune, microbiome, etc.) the BioMapAI had an accuracy of 72.5%. Precision was 0.71 meaning that approximately 7 out of 10 participants predicted to have ME/CFS actually had ME/CFS. On the...

Yes looks like its relatively old data. Patients were diagnosed in 2013–2018 and ethical approval for the study was given in 2016. Weird result. No difference between steroids but almost a total lack of correlations among them in the ME/CFS group. Unfortunately, the authors couldn't give an...

Looks alright for an exploratory analysis like this. They also used a Bayesian analysis where the results of Hanson study (German et al. 2022) on 220 metabolites were used to construct the prior for this study. This means that these metabolites were more or less likely to be statistically...

Some previous studies focused on c4a but results were not entirely consistent: Complement activation in a model of chronic fatigue syndrome - PubMed Unravelling the nature of postexertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome: the role of elastase, complement C4a and...

Here's what this paper says about this:

Wrote a short summary, highlighting some findings: 1) An impressive dataset on ME/CFS was just published by the research team of Ian Lipkin. They tested multiple proteins and metabolites in 56 ME/CFS patients and 51 controls before and after exercise and cytokines in response to mimics of...

Noticed this thread on Twitter which may be useful to merge data with different names for the same gene ID. https://threadreaderapp.com/thread/1948378934967758920.html

Not really convinced by the arguments in this article. Reinfections and PEM from travelling to the site would affect both the placebo and treatment arm, so wouldn't explain the lack of difference between them. Participants were required to have GPCR autoantibodies, similar to the Erlangen...

Social media summary: 1) The Bragée clinic in Sweden published an uncontrolled study to test if intravenous saline improves symptoms in ME/CFS patients with dysautonomia. 2) This is an interesting question because salt solutions that go directly into the vein are often used as a placebo...

Shouldn't the paper also mention that Jeff Wood did/does paid consultancy for people with ME/CFS who think they have CCI as a conflict of interest?

Interferon gamma seems relatively normal compared to controls.

Made some comments about it here:

For other outcome measures they used Tukey's post hoc HSD but for the cognitive testing they used a linear mixed-effects model where it isn't clear if they corrected for multiple comparisons.

The authors used percentage difference scores from visit 1, which often makes the analysis more sensitive to outliers (e.g. participants who had a very low score to start with, could improve by 80-100%). So it would be interesting to check if the results hold up if they used the absolute scores...

I think there is clearly no effect for fatigue or physical functioning, but the scores on the cognitive test (the DANA Brain Vital) do seem to show an improvement in the treatment group compared to controls.

I agree with these comments. There is a risk that this ruling will make things worse because it seems to overvalue the use-case of these tests. I reinforces the idea that you need to have objective abnormalities in order to be able to receive disability (which for some diseases is almost...