I am the Scientific Project Manager for the SEA-AD Consortium. I am posting a community question received via email related to working with SEA-AD data:
All of my questions stem from the “Documentation, Data, and Downloads” section of your website. I accessed the “Donor metadata”, “Quantitative neuropathology summary data (MTG)”, and “SEA-AD cohort metadata summary graphs.”
** Question 1: When comparing Donor Metadata Excel against SEA-AD cohort metadata summary graphs, I found that the Sex bar chart is inconsistent with the Donor Metadata Excel file. The metadata Excel file reports 51 females and 33 males, while the summary graphs represent 52 females and 32 males. If possible, could you clarify this for me or provide the Donor ID that may be flip-flopped?*
** Question 2: In your paper, on page 54, your team mentions “Approximately 25% of ACT autopsies are from people with no MCI or dementia at their last evaluation; roughly 30% meet criteria for MCI, and roughly 45% meet criteria for dementia.” Looking at the Donor Metadata file, I have generated a few questions I am hoping to garner further insight on if possible.*
-
Columns W – AO report checked and unchecked. To confirm, when checked, does this denote a clinical diagnosis of the relevant row?*
-
To confirm, which Donor IDs met the criteria for Alzheimer’s disease? I have generated the following Donor ID’s by filtering column W by checked only.*
-
H20.33.004, H20.33.011, H20.33.015, H20.33.017, H20.33.018, H20.33.020, H20.33.026, H20.33.028, H20.33.029, H20.33.031, H20.33.037, H20.33.045, H20.33.046, H21.33.005, H21.33.007, H21.33.008, H21.33.009, H21.33.027, H21.33.029, H21.33.039, H21.33.042, H21.33.044, H21.33.046*
-
I did not see a column for Mild Cognitive Impairment (MCI). Which Donor IDs met the criteria for MCI?*
-
What was your team’s criteria of MCI?*
-
On the summary graphs file, I noticed the Cognitive Status chart mentioned only 1 patient as “No dementia-MCI”. Is this referring to Donor ID H20.33.036?*
** Question 3: In examination of the SEA-AD cohort metadata summary graphs file, specifically the cognitive status bar graph, what is your team’s definition and Donor ID inclusion criteria for Dementia? To confirm, does this parallel Donor Metadata’s column R?*
Q1: This discrepancy is due to Donor H21.33.021 who identified as female. From this donor’s tissue we identified the presence of a Y chromosome. Therefore, the individual’s gender is female but we list the sex (which we define in the Donor Index, as biological sex, defined by presence of Y chromosome) as male. This is clear in the ABC Atlas here, when filtering by gender = female, and sex = male.
Q2.
A. Yes, when checked this denotes a consensus clinical diagnosis. See our description for Methods of dementia, AD, and MCI detection procedures, and consensus clinical diagnosis description.
B. Yes, these are the donors that had a consensus clinical diagnosis of Alzheimer’s disease. Note that in our own analyses we are not treating the donors as binary - AD vs Control, but as a spectrum of disease (not case-control design). Even so, we want to provide as much clinical information as is available in the donor metadata table. To that effect, you will notice that in Column X are donors where the consensus clinical diagnosis was Alzheimer’s Possible/ Probable. These are donors who had dementia as well. Additionally, about half of the donors have comorbidities, which is common in aged people and those with AD.
C. In column R, we have information related to cognitive status (clinical diagnosis of dementia derived from DSM-IV). The consensus clinical diagnosis for donor H20.33.036 specifically mentioned MCI. We do not have any additional specific MCI diagnosis information for any of the donors.
D. Yes, that is correct. The consensus clinical diagnosis for this donor (H20.33.036) specifically mentioned MCI.
Q3: Yes, that is correct. Cognitive status bar graph matches column R in the donor metadata file. Cognitive status = Dementia is based on the clinical diagnosis of dementia. Also see our descriptions of metadata here.
Hi Dr. Kaplan,
Thank you so much for providing answers to my questions!
Regarding question 2.C, you mentioned that your team does not have additional specific MCI diagnosis information for any donors.
- But if possible, can you provide the Donor ID of the “30% [of autopsies who] meet criteria for MCI” that may be present in your study?
- What were your team’s criteria for MCI?
- If I understand correctly, SEA-AD Donor IDs are a subset of the ACT study.
- Can you please cross-reference this internally to define which of your Donor IDs meets the criteria of MCI? If not, can you please point me in the right direction in accessing " the full medical record is available for research through KPW" (54).
Additionally, I have a new question. On the Quantitative neuropathology summary data (MTG), column I records the value 0 for some Donor IDs. To confirm, does this mean the donor had a 0 µm2 of TDP43 positive grey matter (referencing Seattle Alzheimer’s Disease Brain Cell Atlas (SEA-AD) - Neuropathology Images and quantifications)?
Thank you for all your assistance on this!
Happy to assist.
Is it correct that some of the SEA-AD donors are a subset of the ACT study cohort. 69/84 SEA-AD donors are from the ACT study (ACT vs UW ADRC Clinical Core donors are indicated in the donor metadata sheet in column B).
As I mentioned we do not consider MCI in our analyses (only dementia vs no dementia). The ACT cognitive testing and procedures were designed to detect dementia vs no dementia, and not MCI.
There are some factors why the ACT study is not set up to capture MCI, for example, when the ACT study considers MCI, they find that the CASI cognitive test(which is the primary tool to detect cognitive impairment that might be dementia) is not that great at differentiating between MCI and no MCI. ACT has only rarely gotten into assessing MCI, and I believe the 30% figure is from a small study utilizing ACT participants (unrelated to SEA-AD). You can follow up with ACT with more questions (Contact Us | ACT Study)
Yes, if the value for TDP43 is 0, there was no detectable TDP43 in the analyzed region.
Hello Dr. Kaplan,
Thank you for all of your assistance! I have generated a few more questions I would love to have your assistance on.
Question 1: Column AA of the cohort metadata file is titled “Consensus Clinical Dx (choice=Control).” You mentioned that the SEA-AD does not treat donors as binary, but as a spectrum of diseases.
- Can you help provide context on what criteria assigned the patients as control?
- Similarly, even though these control patients did not have dementia, what made these patients distinct from the remaining no dementia population within the study?
Question 2: I recognize your study uses Thal phases within the cohort metadata file (column BC), and as mentioned in the paper:
“The number of Aβ plaques and pTau+ neurofibrillary tangle bearing neurons in each donor were consistent with traditional staging thresholds for Braak stage and Thal phases, respectively (ExtendedData Fig. 3b)” (page 8).
On the Quantitative neuropathology summary data, there are 6e10 positive objects in grey matter (column CO), a total of 6e10 positive areas in grey matter (column DA), and percent 6e10 positive area Grey matter (DH).
- What was the cutoff for ranges for Aβ plaques in defining each Thal stage relative to 6e10 positive objects/positive area/ percent positve area?
- Similarly, what was your study’s Aβ plaque cutoff?
Question 3: For the Alzheimer’s Disease diagnosis for donors within your study, what is the timeframe from diagnosis to sample collection?
Thank you again for all of your insight!
Hi Kwame, thanks for your questions.
Q1: SEA-AD is not a case-control study, but instead place donors along a spectrum of disease. We do release the consensus clinical diagnosis information derived from our donor tissue sources (UW ADRC and ACT Study), to provide as much info as possible, but we don’t consider this “Control” designation from the consensus conference in our analyses. Donors who do not have dementia, but are not controls per the consensus clinical diagnosis are those with consensus clinical diagnosis = unknown or other. For other there may be some information there from the clinicians to provide context. e.g., there may have other comorbidities. We would suggest instead of the consensus clinical diagnosis, relying more on the cognitive status, neuropath assessments including ADNC, and cognitive scores to draw conclusions. Also note that the NIA-AA criteria for assessing what is AD is based on the overall Alzheimer’s disease neuropathologic change (ADNC) score, which we also include in the metadata file. Those guidelines are that ADNC = not AD is not AD, and that ADNC = low, intermediate, or high are all considered having AD.
Q2: Thal phasing, CERAD scoring, and Braak staging reported are those based on traditional neuropathological assessment, with well-established criteria for neuropathologists for each phase/stage/score. Both traditional staging and the quantitative neuropathological measurement (QNP) you mention are conducted at the University of Washington Biorepository and Integrated Neuropathology laboratory. We compare them in the paper but note that the traditional neuropath assessment and QNP measurements are distinct measurements. We do not base our thal phasing (or other traditional staging) on the QNP measurement.
Q3: For 36/42 donors with dementia we have information on the Age of Dementia Diagnosis (this is available in the donor metadata file and on the donor index table online). Comparing Age of Dementia Diagnosis to Age at Death can provide a timeframe between diagnosis and sample collection at death.