Displaying segregation on Matrix page given sequencing choices associated with LOD score

[selinad]

In coordination with #1549

On Matrix page:

Break segregation into 2 rows (like AR is above it) -- 1st row says: "Candidate gene sequencing" and 2nd row says "Exome/genome or all genes sequenced in linkage region"

Provide counts for each of these (only when score counted) based on pull-down choice in #1549.

Note: if no choice in pull-down the count will not show here, unless you want the pull-down to be required?

The total points column will be the total points from both rows (those two rows are combined into one row for this column) based on the following:

Calculate LOD score based on lower range according to this table:

This means using the "candidate gene sequencing" values in calculating the score -- e.g. summed LOD 0-1.99 gets 0 points summed LOD 2-2.99 gets 0.5 points etc.

@erinriggs @marinadistefano @jennygoldstein @bryceseifert is this correct and where should a note go about the calculated value picking the lower end of the scoring range for segregation (would also love it if someone would be willing to provide text 😁 )

• @wrightmw

[marinadistefano]

I think that if you want to score the segregation evidence (and choose to score it), the pull-down for sequencing method referenced in #1549 should be required so that the tally is displayed accurately. I agree with your display and the choice, as we discussed, to display the lowest scoring option. It might be good to discuss a quick mockup of this on one of the small group calls just to make sure everyone is on the same page.

[marinadistefano]

Text could be something like this: "To aid curators in segregation scoring, the most conservative point values corresponding to the total summed LOD score are provided (candidate gene sequencing only). Please up or downgrade accordingly based on the counts for each sequencing method."

If this is too wordy, let's discuss on the call and shorten it.

[selinad]

Based on 3/26/18 conversation with @marinadistefano @courtneythaxton @tpsneddon --

• Calculate points based on summed LOD score using the candidate sequencing values
• Allow segregation score to be adjusted within range between calculated value and value allowed for exome/genome (this will be run by larger group); do this with 0.25 increments
• Add box to add reason for adjusting score, but do not make it required

[jimmyzhen]

@selinad, @wrightmw,

Any idea on what to do about the existing LOD scores in production data?

[wrightmw]

@marinadistefano @courtneythaxton @ErinRiggs @jennygoldstein

Hey guys, This spreadsheet calculator created by @tpsneddon , did we get the sign off that this is what should be used to calculate the LOD scores?

LODscores2.xlsx

[marinadistefano]

Yes, Heidi said go forth! In our GCI call a few weeks ago with that calculator.

From: Matt W. Wright [mailto:notifications@github.com] Sent: Wednesday, May 16, 2018 5:00 PM To: ClinGen/clincoded clincoded@noreply.github.com Cc: DiStefano, Marina,Ph.D. mdistefano1@bwh.harvard.edu; Mention mention@noreply.github.com Subject: Re: [ClinGen/clincoded] Displaying segregation on Matrix page given sequencing choices associated with LOD score (#1550)

    External Email - Use Caution

@marinadistefanohttps://github.com/marinadistefano @courtneythaxtonhttps://github.com/courtneythaxton @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein

Hey guys, This spreadsheet calculator created by @tpsneddonhttps://github.com/tpsneddon , did we get the sign off that this is what should be used to calculate the LOD scores?

LODscores2.xlsxhttps://github.com/ClinGen/clincoded/files/2010844/LODscores2.xlsx

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-389663786, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgesyBcUellBddqr16J-LFjXAsFZ5Vks5tzJM7gaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

[selinad]

Thx @marinadistefano. We hadn’t seen the file so wanted to confirm this is the correct one

[selinad]

The matrix looks great @jimmyzhen !

Is there a good way for me to check the LOD score calculation? Wow!

[selinad]

@marinadistefano @erinriggs @jennygoldstein @courtneythaxton

To clarify -- if, using Tam's calculator, the calculated points is 2.1578.., I want to confirm that we don't use 2.16 as the points, be we round to nearest 0.25, which would be 2.25 (see columns below from Tam's spreadsheet):

Thanks!

[marinadistefano]

Yeah I agree. I feel like that’s the most practical thing to do Selina. I’m good with that if others agree.

Sent from my iPhone

On Jun 14, 2018, at 8:39 PM, selinad notifications@github.com<mailto:notifications@github.com> wrote:

    External Email - Use Caution

@marinadistefanohttps://github.com/marinadistefano @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein @courtneythaxtonhttps://github.com/courtneythaxton

To clarify -- if, using Tam's calculator, the calculated points is 2.1578.., I want to confirm that we don't use 2.16 as the points, be we round to nearest 0.25, which would be 2.25 (see columns below from Tam's spreadsheet):

[image]https://user-images.githubusercontent.com/11320314/41444678-7a512944-6ff9-11e8-981c-db6c8f8495ca.png [image]https://user-images.githubusercontent.com/11320314/41444654-4f7dc97a-6ff9-11e8-93f5-cfa144b69305.png

Thanks!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-397479601, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgerzSpOazT0HN3UtL47iWVy5BJY3Jks5t8wI-gaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

[ErinRiggs]

That sounds fine to me.

[selinad]

ok, thanks, @marinadistefano , @ErinRiggs -- we wanted to make sure since there are 0.1 increments for probands and this would involve rounding up by more than that amount. If more time is needed to check with leadership, that is fine as it will be in the release after this next one.

@jennygoldstein -- can you please make certain all of this scoring is put in the SOP? -- when we display this info on the website, it will need to have a matching SOP. thank you!

[courtneythaxton]

Just a question... what happened when you. are between to 0.25 increments. for instance if your LOD happens to be 2.375 (between 2.25 and 2.5), do we always round up? Or even more so, if we are at 2.875, does this round up to 3, which would increase the total points given from 0.5-1pt (even though the original LOD score would deem it in the range for 0.5 pts)? I wonder if in the upper range of the tenths place it should only round up to 0.99, so as not to peak it over to a new range arbitrarily. Just a thought

[marinadistefano]

We have a GCI call next week. Maybe we can just ask the PIs what they think quickly at the beginning?

From: courtneythaxton [mailto:notifications@github.com] Sent: Friday, June 15, 2018 10:59 AM To: ClinGen/clincoded clincoded@noreply.github.com Cc: DiStefano, Marina,Ph.D. mdistefano1@bwh.harvard.edu; Mention mention@noreply.github.com Subject: Re: [ClinGen/clincoded] Displaying segregation on Matrix page given sequencing choices associated with LOD score (#1550)

    External Email - Use Caution

Just a question... what happened when you. are between to 0.25 increments. for instance if your LOD happens to be 2.375 (between 2.25 and 2.5), do we always round up? Or even more so, if we are at 2.875, does this round up to 3, which would increase the total points given from 0.5-1pt (even though the original LOD score would deem it in the range for 0.5 pts)? I wonder if in the upper range of the tenths place it should only round up to 0.99, so as not to peak it over to a new range arbitrarily. Just a thought

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-397648917, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgeteYKBlFIVSrSdNqSU4936BFiJ_mks5t88u8gaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

[courtneythaxton]

I agree it might be a good idea to ask them. Typically we tend to lean on the conservative side, so as not to over-inflate scores. Rounded up could definitely cause an inflation in points given for segregation.

courtney

On Jun 15, 2018, at 11:57 AM, marinadistefano notifications@github.com<mailto:notifications@github.com> wrote:

We have a GCI call next week. Maybe we can just ask the PIs what they think quickly at the beginning?

From: courtneythaxton [mailto:notifications@github.com] Sent: Friday, June 15, 2018 10:59 AM To: ClinGen/clincoded clincoded@noreply.github.com<mailto:clincoded@noreply.github.com> Cc: DiStefano, Marina,Ph.D. mdistefano1@bwh.harvard.edu<mailto:mdistefano1@bwh.harvard.edu>; Mention mention@noreply.github.com<mailto:mention@noreply.github.com> Subject: Re: [ClinGen/clincoded] Displaying segregation on Matrix page given sequencing choices associated with LOD score (#1550)

External Email - Use Caution

Just a question... what happened when you. are between to 0.25 increments. for instance if your LOD happens to be 2.375 (between 2.25 and 2.5), do we always round up? Or even more so, if we are at 2.875, does this round up to 3, which would increase the total points given from 0.5-1pt (even though the original LOD score would deem it in the range for 0.5 pts)? I wonder if in the upper range of the tenths place it should only round up to 0.99, so as not to peak it over to a new range arbitrarily. Just a thought

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-397648917, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgeteYKBlFIVSrSdNqSU4936BFiJ_mks5t88u8gaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-397665733, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVRHpvR6okcZnOArybVUKH4eKxyJKks5t89lrgaJpZM4QfiqI.

[ErinRiggs]

Can we discuss this at the beginning of our GCI call on 7/19? If possible, could we please see a mock-up or workflow of how the updated segregation scoring will be enabled? That will help us decide these point increments.

[selinad]

@ErinRiggs Jimmy has implemented this in dev according to the tickets ( #1549 and #1550) -- we should be able to show you in an instance on the call (will confirm with Jimmy when he is back). We are ready to go once the rounding up question has been decided and the website team is able to display as required.

[ErinRiggs]

Great! We talked about this a bit today and had a few questions that would help us decide (which viewing your instance may be able to answer): Will the GCI display for you the suggested score based on the result of Tam's calculation (say, for the sake of example, it is the 2.15789...from the screenshot above), then have the curator select the actual points they want to award from a pulldown? Or will the GCI assign a score (based on whatever rounding we decide) with no option to adjust? Must we round up to the nearest 0.25?

It's been awhile since we last discussed this, so forgive us if some of these questions have already been addressed. Just want to make sure we are prepared with all information the PIs need to make a decision on 7/19!

[selinad]

Hi @ErinRiggs -- my memory of how it works (the instance is down right now) is it calculates the LOD score using Tam's algorithm and gives the score (points) to 2 decimal points (e.g 2.16 points). We were not planning to allow curators to change the points -- just as they do now, curators could make adjustments in the section where Classifications are modified. Allowing curators to adjust would introduce a fair amount of complexity.

And, you do not need to round up -- Tam had included a rounded up value, which is why we wanted to check. Right now, we do include points to the hundredths decimal point, but only 0 and 5 are allowed in that position (e.g. 0.25 for a proband). If we don't round up, we will have things like 2.16.

One thing I just noticed -- I think @courtneythaxton was talking about rounding up LOD scores (?), but this is about rounding up points (not rounding the LOD score up by 0.25). It's possible I'm confused, though!

When the instance is back up again, we could send you the URL so you can play around with it, if that would help.

[courtneythaxton]

Hi @selinad and @ErinRiggs,

Initially I was confused about the calculator and where the rouding up would occur. After yesterday’s call I now understand better the situation. I think it might be good to go over the logisitics of allowing a points adjustment though.

Also I had a few questions.

1) Even with the rounding up of the final allotted segregation points, each LOD score/ range has a cap on the total number of points you could receive. For instance a LOD score of 2-2.99 can receive theoretically between 0.5 points at the lowest end (although starting at 0 might be best), but the maximum is 1pt. Does the GCI take this into account? Would we want it to, or just allow for a final adjustment by the curators?

2) With Tam’s calculator and algorithm, he has adjusted the points distribution based on two or more segregation scores. Does Tam’s calculator also work if you have only one segregation to include? I ask this because off hand, it appears that if you have one segregation count from say a exome sequencing with a LOD score of 3, this could get 2 pt. Now if we had another situation in which we have two families with segregation, one exome sequencing with a LOD of 3 and one candidate with a LOD of 1.2, combined LOD= 4.2, this presumably would get less than 2, as the points range for LOD 3-4.99 is 1-2 pts total. Here, I worry that the addition of an extra piece of evidence is actually scoring the total points down.

• What would we normally suggest in this scenario? Does Tam’s calculator/ algorithm suggest the points should be for case of one segregation with exome sequencing above? Would it score down from the maximum 2 pts?

I would definitely appreciate being able to run through the test instance if possible.

Thanks! Courtney

Courtney Thaxton, Ph.D. Biocurator Dept. of Genetics 120 Mason Farm Road 5100 B, Genetic Medicine Building CB#7264 University of North Carolina, Chapel Hill, NC 27599 919-843-3549

On Jul 9, 2018, at 4:43 PM, selinad notifications@github.com<mailto:notifications@github.com> wrote:

Hi @ErinRiggshttps://github.com/ErinRiggs -- my memory of how it works (the instance is down right now) is it calculates the LOD score using Tam's algorithm and gives the score (points) to 2 decimal points (e.g 2.16 points). We were not planning to allow curators to change the points -- just as they do now, curators could make adjustments in the section where Classifications are modified. Allowing curators to adjust would introduce a fair amount of complexity.

And, you do not need to round up -- Tam had included a rounded up value, which is why we wanted to check. Right now, we do include points to the hundredths decimal point, but only 0 and 5 are allowed in that position (e.g. 0.25 for a proband). If we don't round up, we will have things like 2.16.

One thing I just noticed -- I think @courtneythaxtonhttps://github.com/courtneythaxton was talking about rounding up LOD scores (?), but this is about rounding up points after the LOD score is rounded to the nearest hundredth (not rounding the LOD score up by 0.25). It's possible I'm confused, though!

When the instance is back up again, we could send you the URL so you can play around with it, if that would help.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-403613961, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVdOO0EK4pGQ4Vt_IL6gDj4L0lXmgks5uE8COgaJpZM4QfiqI.

[wrightmw]

@courtneythaxton Hi Courtney, so following on from the discussion just now. There were a few different suggested options: 1) To implement the scoring as per Tam's calculator (simplest to implement because Jimmy has already done the work on this one) 2) Add in a piece of logic. Whereby Tam's calculator is used by default but if a single piece of evidence has a greater score than the score derived from Tam's calculator, then that should be used instead. 3) Your third option. Can you please clarify the logic for your proposal?

[wrightmw]

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

Here is the top of the Classification matrix that Jimmy created based on your specs. As you can see, he has split the segregation into two rows: "Candidate ..." and "Exome/genome...". The "Total Points" cell for Segregation data just includes the total points based on Tam's calculator and the summed LOD score in parentheses:

On the call today, it seemed like you would also like to see the total Candidate score and the total Exome/genome score within this cell in the table. Could you please specify how you'd like these to be shown in this Classification matrix?

[courtneythaxton]

HI Matt,

Thanks for the email, and questions. The last specification was as you mentioned, I think that Sharon requested more information on the segregation be displayed. I was thinking something along the lines of this (see attached pdf). I mocked it up on an old spreadsheet as that was handier and quicker. I tried to make all the adjustments like the newer display but please forgive me if I missed something. I think the gist of it, is that they want to see the summed LOD score for each type of evidence with its category. Then you can show the total LOD score with the points distribution in the Total points. The count would also reflect the total pieces of segregation evidence as always.

This is just my quick idea, if anyone else has some examples please send along!

Courtney

Courtney Thaxton, Ph.D. Biocurator Dept. of Genetics 120 Mason Farm Road 5100 B, Genetic Medicine Building CB#7264 University of North Carolina, Chapel Hill, NC 27599 919-843-3549

On Jul 19, 2018, at 6:12 PM, Matt W. Wright notifications@github.com<mailto:notifications@github.com> wrote:

@marinadistefanohttps://github.com/marinadistefano @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein @courtneythaxtonhttps://github.com/courtneythaxton

Here is the top of the Classification matrix that Jimmy created based on your specs. As you can see, he has split the segregation into two rows: "Candidate ..." and "Exome/genome...". The "Total Points" cell for Segregation data just includes the total points based on Tam's calculator and the summed LOD score in parentheses:

[screen shot 2018-07-19 at 2 48 49 pm]https://user-images.githubusercontent.com/15131169/42972845-c279030e-8b65-11e8-80d9-070afd9bda90.png

On the call today, it seemed like you would also like to see the total Candidate score and the total Exome/genome within this cell in the table. Could you please specify how you'd like these to be shown in this Classification matrix?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-406430509, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVSON2PhZdQ4vJe_P1XAd6UfnIeEEks5uIQRogaJpZM4QfiqI.

Segregation display (1).pdf

[wrightmw]

Courtney's pdf image:

[wrightmw]

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

I've pasted Courtney's proposal from email thread below:

Formula:

(Maximum total points for Total LOD range) - [(maximum points- minimum points in range)*(number of candidate gene sequencing evidence count/ total evidence count)]= adjusted segregation score

Again, the use of this equation would still need the logic built in that if any one piece of evidence for a LOD range meets the maximum score on that this formula is not used and superceded by giving the max points in that range. But for the remaining evidence of multiple lines of evidence that do not meet the max of min score for the total LOD score range it should work.

[wrightmw]

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

From the emails today, it looks like we have a resolution. Which I think is that Tam's calculator is going to be used but we need to add in additional logic such that if a single piece of evidence has a greater score than the score derived from Tam's calculator, then that should be used instead. Can you confirm that is correct? And if so, can one of you please specify this logic? For instance, if there are multiple pieces of evidence that have a higher score than Tam's calculator I assume you would use the highest of these scores? It would be good if you could provide several examples of how the logic might work in different scenarios. This will help Jimmy to implement it, and will avoid us having to come back to you later with further questions.

Also, one last clarification. I think we agreed to go with rounding Tam's calculator to 0.1 instead of 0.25. Is that correct?

[marinadistefano]

Yes. That resolution is correct. And we did agree on the 0.1 rounding for Tam’s calculator. Whenever you get a chance, Courtney, could you send along some examples for the logic? You’ve worked out so many I figured that it would be quick for you.

From: Matt W. Wright [mailto:notifications@github.com] Sent: Friday, July 27, 2018 12:08 PM To: ClinGen/clincoded clincoded@noreply.github.com Cc: DiStefano, Marina,Ph.D. mdistefano1@bwh.harvard.edu; Mention mention@noreply.github.com Subject: Re: [ClinGen/clincoded] Displaying segregation on Matrix page given sequencing choices associated with LOD score (#1550)

    External Email - Use Caution

@marinadistefanohttps://github.com/marinadistefano @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein @courtneythaxtonhttps://github.com/courtneythaxton

From the emails today, it looks like we have a resolution. Which I think is that Tam's calculator is going to be used but we need to add in additional logic such that if a single piece of evidence has a greater score than the score derived from Tam's calculator, then that should be used instead. Can you confirm that is correct? And if so, can one of you please specify this logic? For instance, if there are multiple pieces of evidence that have a higher score than Tam's calculator I assume you would use the highest of these scores? It would be good if you could provide several examples of how the logic might work in different scenarios. This will help Jimmy to implement it, and will avoid us having to come back to you later with further questions.

Also, one last clarification. I think we agreed to go with rounding Tam's calculator to 0.1 instead of 0.25. Is that correct?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-408465305, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgelAEGhmDBPwNd1Go-nRLTk3a9kr_ks5uKzragaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

[courtneythaxton]

Hi Matt,

Yes it is my understanding that the logic we want to code, is that if you have more than one piece of evidence for both segregation categories (i.e. candidate and exam/genome) and a single LOD score for exome/genome meets the requirement for the full points in that LOD scoring range, that the addition of the LOD for the second piece of evidence (i.e. candidate) should not reduce the points (use the calculator), if it does not move the total LOD into a new LOD scoring range than the single LOD of the exome/genome.

Example: 1 count exam genome LOD=3.1, 1 count of candidate LOD=1.2, 2 counts of total evidence for a summed LOD= 4.2. The LOD scoring range is for the 3-4.99 range where candidate gets 1 pt, exam/genome can get up to 2 points. In this example the single piece of exome/genome alone is within the scoring range of 3-4.99, thus would get 2 points. The addition of the candidate LOD of 1.2, does not move the total LOD out of the scoring range of 3-4.99, therefore the calculator should not be used. If the calculator is used it will produce a score of 1.75, which is lower than the 2 points the single piece of exome/gnome LOD=3.1 would achieve on its own.

Other example: 1 count genome LOD= 5, 1 count candidate LOD=1.2, 1 count candidate LOD= 2, total genomeLOD= 5, total candidate LOD= 3.2.

The points should reflect the maximum for the genome of 3 as the one piece of genome along had a LOD score of 5 for one singular piece of evidence.

Where we may not want to use this logic?: . 1 count exome LOD= 2.5, 1 count genome LOD= 2.5, 1 count candidate LOD= 1.2 one count candidate LOD=3. Total exome LOD= 5(2 evidence), total candidate LOD= 4.2.

Here I may suggest using the calculator, as the strongest piece of single evidence is the one piece of candidate data with LOD= 3. And because the exome LOD comes from two pieces of evidence and the summed LODs from each evidence category are close (exome = 5, candidate = 4.2) I would want to use the calculator. I would be interested what others think here? @ErinRiggs @marinadistefano @jennygoldstein

Hopefully this is clear. This will only really happen in the situations where one singular LOD score of exome/genome segregation meets the minimum the requirement of scoring within the same scoring range as the summed LOD score including additional evidence (i.e. candidate segregation).

As far as the rounding, yes , it is my understanding that we would like the rounding to occur to the nearest tenths place (0.1) versus the 0.25 increment.

[courtneythaxton]

I updated the GitHub thread with some examples. One example I have as one where we might not want to use the logic but the calculator, and tagged everyone to give their opinion, that would be great!

We’re moving aLOD on this one, yay!

Courtney Thaxton, Ph.D. Biocurator Dept. of Genetics 120 Mason Farm Road 5100 B, Genetic Medicine Building CB#7264 University of North Carolina, Chapel Hill, NC 27599 919-843-3549

On Jul 27, 2018, at 12:54 PM, marinadistefano notifications@github.com<mailto:notifications@github.com> wrote:

Yes. That resolution is correct. And we did agree on the 0.1 rounding for Tam’s calculator. Whenever you get a chance, Courtney, could you send along some examples for the logic? You’ve worked out so many I figured that it would be quick for you.

From: Matt W. Wright [mailto:notifications@github.com] Sent: Friday, July 27, 2018 12:08 PM To: ClinGen/clincoded clincoded@noreply.github.com<mailto:clincoded@noreply.github.com> Cc: DiStefano, Marina,Ph.D. mdistefano1@bwh.harvard.edu<mailto:mdistefano1@bwh.harvard.edu>; Mention mention@noreply.github.com<mailto:mention@noreply.github.com> Subject: Re: [ClinGen/clincoded] Displaying segregation on Matrix page given sequencing choices associated with LOD score (#1550)

External Email - Use Caution

@marinadistefanohttps://github.com/marinadistefano @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein @courtneythaxtonhttps://github.com/courtneythaxton

From the emails today, it looks like we have a resolution. Which I think is that Tam's calculator is going to be used but we need to add in additional logic such that if a single piece of evidence has a greater score than the score derived from Tam's calculator, then that should be used instead. Can you confirm that is correct? And if so, can one of you please specify this logic? For instance, if there are multiple pieces of evidence that have a higher score than Tam's calculator I assume you would use the highest of these scores? It would be good if you could provide several examples of how the logic might work in different scenarios. This will help Jimmy to implement it, and will avoid us having to come back to you later with further questions.

Also, one last clarification. I think we agreed to go with rounding Tam's calculator to 0.1 instead of 0.25. Is that correct?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-408465305, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcqgelAEGhmDBPwNd1Go-nRLTk3a9kr_ks5uKzragaJpZM4QfiqI.

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-408477568, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVauU_mwnUPCd7bMo8mAsr6e6_N49ks5uK0XUgaJpZM4QfiqI.

[courtneythaxton]

Hi All,

I have mocked up the segregation scoring display for the GCI. I have also added numbers to provide an example.

Let me know if I need to make any adjustments. @ErinRiggs @selinad @wrightmw @marinadistefano @jennygoldstein Segregation display v2.pdf

Thanks so much @courtneythaxton ! Here it is:

[jennygoldstein]

Thanks Courtney. It looks good. A couple of thoughts: Should we label the "LOD" column "total LOD" instead. And do we need the "Count" column as it will give the same number as the third row count. What do you all think?

[marinadistefano]

I agree that this also looks good. I think the "total summed evidence" label hints that the LOD is summed, but maybe it's not entirely clear? I agree that the count column is a bit redundant, but it exists for all of the other evidence types. I guess we could gray it out or get rid of the count section in the total summed evidence?

[courtneythaxton]

Alright everyone I have a version 3 of the segregation data display on the GCI. this one grays the "count" column shared over all of the cells for genetic evidence.

Let me know what you think: @marinadistefano @ErinRiggs @jennygoldstein @wrightmw @selinad Segregation display v3.pdf

[marinadistefano]

That looks great to me! Thank you for doing all of this Courtney.

[wrightmw]

How about moving all the the Counts over to the Counts column...something like this?:

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

[courtneythaxton]

I had it like that originally but wasn’t sure if the addition of the third row of “total summed evidence” made it confusing. While we know it is two total counts of evidence could it be confused for 4 counts.

I can go either way.

I do wonder though why only the genetic evidence categories show counts of evidence and not experimental? Maybe this would be a separate ticket :)

@marinadistefan @erinriggs @jennygoldstein

Sent from my iPhone

On Aug 3, 2018, at 8:17 PM, Matt W. Wright notifications@github.com<mailto:notifications@github.com> wrote:

How about moving all the the Counts over to the Counts column...something like this?:

[screen shot 2018-08-03 at 5 16 01 pm]https://user-images.githubusercontent.com/15131169/43670474-121ab286-9741-11e8-8328-0b3540cc91ce.png

@marinadistefanohttps://github.com/marinadistefano @ErinRiggshttps://github.com/ErinRiggs @jennygoldsteinhttps://github.com/jennygoldstein @courtneythaxtonhttps://github.com/courtneythaxton

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-410406314, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVahMxNHh6U5_bChzW7KQfsa5n9Iyks5uNOgpgaJpZM4QfiqI.

[marinadistefano]

Sorry for the delay in reply. I'm fine with either of these mock ups.

[wrightmw]

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

We're currently implementing the logic and we have a few questions. Looking at your examples @courtneythaxton

---

Example 1: Exome/genome LOD: 3.1 Candidate LOD: 1.2 Total LOD: 4.2 (Range 3-4.99)

Calculator score is 1.72 Exome score is 2 (based on range)

So in this case we would use the Exome score which is 2.0

---

Example 2: Genome LOD: 5 Candidate LOD: 3.2 Total LOD: 8.2 (Range >5)

Calculator score is 2.42 Exome score is 3.0

So in this case we could use the Exome score which is 3.0

---

Is this all correct?

[courtneythaxton]

Hi @wrightmw ,

Yes for both of these examples you are correct for the score, and interpretation of the logic.

[wrightmw]

Just checking a couple more scenarios with you guys to make sure we have our calculator working correctly. Please define what you think the total score should be for each:

---

Example 3: Candidate LOD: 5.66 Exome LOD: 2 Total Summed LOD: 7.66 Calculator Score is 1.89

---

Example 4: Candidate LOD: 0.5 Exome LOD: 1.2 Total Summed LOD: 1.7 Calculator Score is 0

---

Example 5: Candidate LOD: 6.66 Exome LOD: 3.3 Total Summed LOD: 9.96 Calculator Score is 1.99

---

@marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

[courtneythaxton]

Hi Matt,

For example 3, I think the final score should be rounded to the nearest 0.1, so it should be 1.9 (although you do have the correct calculation before the rounding)

Example 4: It is true that this should not receive a score. Interestingly I would like to point out that for our restrictions on the number of segregations or affected required for AD or AR MOI, that the lowest LOD score we can actually consider to use for points is 1.2. It might be good to consider logic that if any single LOD score is less than 1.2 it it not used for the calculator. thought?**

Example 5: Again with the rounding it should be 2 pts (but otherwise correct) @marinadistefano @ErinRiggs @jennygoldstein @courtneythaxton

[courtneythaxton]

Okay all,

I have another mockup based on the feedback from the Gene Curation Working Group call today.

Maybe show this one on the VCI call to get more opinion @marinadistefano @ErinRiggs @jennygoldstein @wrightmw @jimmyzhen

Segregation display v7.pdf

[courtneythaxton]

Okay with the new calculator form Tam, I can see that e has actually coded in the logic we went over to the excel calculator, which is great.

It did bring me to the question I asked earlier of another potential logic for consideration. This logic might not be needed all the time, as long as curators are clear of the segregation rules, which families can be included and are diligent in clicking the button saying "do not include segregation" but there is a potential to get around this with logic.

The logic being, that at this time for both AD and AR, the minimum required segregations for each ends up having a LOD of 1.2. So the minimum LOD score for all segregation evidence required to even begin to tally for segregation scoring is 1.2.

So would it be possible to include logic that if any single LOD score if below 1.2, it is not included in the calculator or segregation points tally?

@marinadistefano @ErinRiggs @wrightmw @jimmyzhen @jennygoldstein

[wrightmw]

@courtneythaxton Is there a new version of Tam's calculator? When was this new version created? Can you please post it in this ticket? Thanks.

[courtneythaxton]

I just received it this morning, here it is

So far it has worked well for all examples, but again if you enter a LOD score that actually does not meet the minimum LOD (i.e, any single LOD score below 1.2) it will still tally.

LODscores3.xlsx

[wrightmw]

Thanks @courtneythaxton .... what has changed in this new calculator?

[courtneythaxton]

Tam actually coded in the logic that if the exam summed LOD meets the minimum requirement for the full points that addition of the candidate does not drop the points.

So essentially he coded in the logic in the excel calculator :)

[jimmyzhen]

@wrightmw, @selinad,

I have spun up a new instance for your review on the latest round of changes on the segregation: https://1549-jz-lod-score.demo.clinicalgenome.org

Please verify the following changes on this instance:

1. The "Summed LOD" scores are rounded to 2 decimals in the classification matrix.
2. The new formatting of the table cells for the segregation in the classification matrix.

Let me know if you have any questions. Thanks.

[courtneythaxton]

Hi All,

As requested on the VCI call I have generated a mockup of the segregation display based on recommendations from Scott G. on the website side and the PIs. You may find some columns that are meant for the website display, but I wanted to replicate the base that I was given.

Let me know if there is anything else that needs to be altered. @marinadistefano @wrightmw @ErinRiggs @jennygoldstein @jimmyzhen @selinad

Courtney @courtneythaxton Segregation display v8.pdf

[courtneythaxton]

This looks great Jimmy!

The LOD scores that I tried appropriately only returned to the hundreths decimal (0.01.

The segregation points appropriately rounded to the nearest tenths place (0.1), and the display is spot on.

Really nice and easy to understand the information.

Courtney

Courtney Thaxton, Ph.D. ClinGen Biocurator Berg Lab, UNC Dept. of Genetics 120 Mason Farm Road 5100 B, Genetic Medicine Building CB#7264 University of North Carolina, Chapel Hill, NC 27599 919-843-3549

On Aug 10, 2018, at 11:06 AM, Jimmy Zhen notifications@github.com<mailto:notifications@github.com> wrote:

@wrightmwhttps://github.com/wrightmw, @selinadhttps://github.com/selinad,

I have spun up a new instance for your review on the latest round of changes on the segregation: https://1549-jz-lod-score.demo.clinicalgenome.orghttps://1549-jz-lod-score.demo.clinicalgenome.org/

Please verify the following changes on this instance:

1. The "Summed LOD" scores are rounded to 2 decimals in the classification matrix.
2. The new formatting of the table cells for the segregation in the classification matrix.

Let me know if you have any questions. Thanks.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ClinGen/clincoded/issues/1550#issuecomment-412111206, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AfBeVQnzoD-VneZQs2GcLFREsLdBtf17ks5uPaGEgaJpZM4QfiqI.