database training academy review

good afternoon and welcome to the ncbi webinar on using osiris. my name is ben busby, and i'm the genomics outreach coordinator for ncbi. i'm joined today by george riley, rob goor, doug hoffman, and tao tao, all staff scientists here at ncbi, many of whom have worked on the osiris project. we still have a few people coming in the door, and we're going to start in just a couple of minutes. while people are still coming in the door, i wanted to show those of you who are more biologically oriented that ncbi, in fact, has a series of webinars. and we'll be talking about our api next month, as well as downloading genomic sequences the month after. this webinar, once we get the closed caption straightened out in about two weeks, will go in the archives.

and you'll be able to refer back to the video of this webinar through the archived webinars tab. to get to this page, you can simply google "ncbi webinars." asking questions in this webinar -- please type them into the questions module, which you'll see in the go to webinar control interface on the right side of your screen. like i said, type in questions. quick technical questions, we'll just answer via chat. and then more complex questions, what we'll do is we'll ask george at the end and make sure he is straightened out with those. now i'm going to turn the presentation over to george riley,

who is going to tell you about the osiris analysis summary. hi, i'm george riley. i'm a staff scientist at ncbi. i have experience in biology and forensic dna analysis. i was a forensic dna analyst and dna laboratory director for ten years and additional experience in laboratory quality systems. today, i'd like to talk to you about the osiris str analysis software and do a short demonstration. i'm going to start the slide show at this point. okay, so today we'll talk about osiris software.

first off, osiris is the open source independent review and interpretation software. it's public domain software. i'd like to point out that the commercial manufacturers and products that we mention today are only mentioned incidentally and don't imply endorsement by ncbi or nih. osiris was conceived in response to recommendations from the world trade center victim identification project. it was first tested in a pre-release mode in the hurricane katrina victim identification, on which i worked, and was further developed in collaboration with federal and state laboratories and the national institutions of standards and technology. osiris is designed to give rapid str analysis with the identification of artifacts.

it has been accepted by ndis as an expert system for doing codis convicted offender analysis. it is public domain software, so it's freely available for download from the ncbi osiris home page. i would point out, please do sign up for the announcements on that page so that you'll get announcements of a new version which is coming out soon. today, we'll be discussing version 2.3; but version 2.4 will be available soon. the source code is publically available on github as well. osiris runs on windows xp, windows 7. i'm told it runs on windows 8, but we have not excessively tested that; and it runs on the macintosh as well.

it analyzes data in the .fsa and .hid file formats, which are produced on the applied biosystem's 310, 3100 series, 3500, and 3700 series genetic analyzers. it's not limited by the number of str kit dye colors. so if somebody invents a kit that has eight colors or more, it will work on osiris. also, if it's run on the network, it's not limited to a certain number of users. it's very fast; at a quarter of a second per sample, osiris will analyze an entire 96 well plate in under 30 seconds. osiris works differently than most of the other str analysis software. it has a mathematical basis. it works by fitting mathematical curves to the raw data peaks.

the alleles are fitted to double-gaussian curves. and the artifacts -- such as spikes, blobs, split peaks, which we call "craters" -- have other mathematical signatures that control the alleles and the ladder. the time is mapped to the base pair sizes using a cubic spline. and when analyzing samples, osiris does something a little bit different than other software. what it does is it takes each sample and compares the sample ils to the ladder ils to ensure the best sample-to-ladder ils fit. that means that you get less shifting when you do your analysis. then you do peak identification and artifact identification.

osiris introduces some new quality metrics, which we'll talk about a little later. and it also does something interesting in that we set different artifact priorities. it is critical artifacts; these are artifacts which indicate actual issues in quality that require review and non-critical artifacts which the laboratory has decided do not need review. osiris introduces a number of concepts including fit, the goodness of fit being a mathematical curve to the raw data peak. this is one of the in-quality metrics. another quality metric is residual. this indicates the distance the sample peak shifts away from the ladder peak center.

and there is primary pull-up; this is the peak that causes pull-up. and finally, peak has restricted priority; these are non-critical artifacts which fall below a threshold -- either the analysis threshold, stutter threshold, et cetera. and those do not get real cause. so the take-home is that osiris is a fast and flexible str analysis. it gives you automatable profile qc with artifact identification. it has a very flexible export so that you can actually get your data into whatever tables or (technical difficulty) you would like.

it's validated as an expert system for codis samples by the illinois state police indexing laboratory. and it's been validated for casework by the u.s. army crime lab. basically what it's designed to do is increase accuracy and reduce the amount of time that you spend looking at your profiles. when you are looking at osiris, there are a couple of things that can be very helpful. the help menu and f1 will allow you to access the users guide. the table of contents and adobe reader bookmarks and control f can be used when you're looking at that users guide to ensure that you can find what you're looking for. finally, i definitely would look at the faq and the troubleshooting guide,

which are appendices of the user guide because that has information that will help you with a lot of the problems that you might run into. you can, in the program itself, hover over buttons for tool tips telling what the buttons do. and as a sneaky way of getting your data out quick and dirty, you can actually -- when you're in the table view -- you can press control c, or command c on the mac, and you can then paste your data, your allele calls straight into a print spreadsheet. when you're installing osiris, it does not require administrative privileges. however, you do have to have write privileges in the installation directory that you put it. so for example, on your documents directory or on the desktop directory you have write privileges.

these are good places to put it. alternatively, if you put it into something like the programs file and, as you see, you can change the write privileges with the installation directory. so setting up for forensic casework, there are a couple of things that i would suggest strongly. one of them is doing normalization; this is for baseline subtraction. and there are also differential filter settings for mixture calculations. so if you're doing mixture calculations using software, you can actually set osiris up so that it will treat your reference samples, your knowns, differently from your mixture samples.

and your applied filters will not apply filters to your mixture calculation, your mixture samples, as you choose. normalization basically for baseline subtraction, as you can see here. this has been normalized. it's had baseline subtracted from the raw data. so the light pink line here, you can see that is the baseline that was subtracted from the raw data. you can see the raw data here with the noise included gives you a very flat baseline. that's important because this small stutter peak would have been probably half again as high if it had had that baseline included in its peak height. so when you start osiris, the first thing that you'll see is you'll see the logo.

and if you have done analysis before, you'll see the recently viewed files window. if you have not done analysis before, you won't see this until you've saved some files. you can type into the search bar to restrict the files. and that way once you have a lot of files, you can actually find what you're looking for. the clean up list button will delete the old deleted files. when you start your new analysis, you go to file/new analysis, or you can press control n. this opens up the new analysis window. first you put it in your input directory, and then you put in your output directory. i would suggest when you put your output directory that you choose something that isn't your installation directory

so you don't have that confused with a large number of files from analyses. then you choose your operating procedure name. the operating procedure is what we call the lab settings for a particular kit definition. choose your internal lims standard. and then if these are not grade-out, you can actually modify the analysis threshold and the ladder and ils thresholds as well. finally, you press okay; your analysis runs. when you do a single analysis, osiris will -- actually if it succeeds, the analysis will open up into the table view. otherwise, a batch window will open.

osiris will analyze all of the subdirectories and select a directory. so if you put multiple folders in, you can analyze them all at one time. if you've analyzed multiple directories, it will tell you the status -- whether they've failed or whether they've completed. if your analysis fails, there are a number of reasons. the most common ones -- my favorite one, this i do myself -- is i choose the wrong operating procedure. you have to choose the right kit definition, or it won't work. similarly, if you choose the wrong ils, if there is no ladder -- you need to have a ladder present in your analysis for the run to succeed.

if the ladder ils peaks are below threshold or if some of the peaks are missing -- such as if they're not collected. finally, if the ladder peaks are below threshold or if there are no .fsa's; these are all causes of failures. once your analysis is complete, what you do is you select the analysis that you want to look at; you click the view selection button. if your analysis fails, you can actually select that and view selections so that you can troubleshoot the ladder to check to see if all of your peaks are there. there is also a details button, which will give you other troubleshooting details. and if you need to ask us questions, the details there can sometimes be helpful so that we can give you some idea of what to do.

you can also reanalyze your selection by changing the thresholds. so if your ils, if some of those peaks were below threshold, you can actually change the threshold and rerun your analysis. when it first opens up, what you see is the table view. this is a little bit different then what most people are used to. there is also a graph view, the classic sort of stacked graphs -- electropherograms -- that people are used to seeing. in the table view, you have basically three panels. the top panel is your table. the center panel gives you a preview of whichever sample of locus you want to look at in that table.

and the bottom panel gives you notifications of issues and problems. the preview button at the upper left will actually allow you to turn that preview pane on and off to give yourself some more screen real estate if you need it for the table. when you're looking at the table, there are basically a number of different columns. there are the columns for loci. there is a column for channels that may have issues; for the internal marker that may have issues; and the sample column, which will give you either the green checkmark or the red x. the green checkmark means the sample has passed with no issues. the red x means that there are still remaining issues that need to be addressed.

the cells themselves in the loci will be either white for having passed. this is blank because the ladder has too many alleles. we don't put the alleles calls unit. the yellow cells are indicative of a locus that has passed in a sample that has other issues. the red cells show loci that have problems. when you're looking at the preview, if you want to take a look at a particular locus, you click the locus. that will zoom the preview in to the locus in question. down in the bottom right panel, then, you'll see notices of problems. in this particular case, the locus has a peak with laser off scale at that location.

you can see that very small peak there, which tends to be indicative of pull-up. in the left-hand panel when you're in this view, you will see information about the peaks. what you can do also is you can click a sample. for example, click the positive control here. and at that point, in the lower left-hand, you'll see all the alleles, which is not particularly helpful. but in the lower left-hand panel, you'll also see helpful links to accomplish editing of the various issues in that sample. in the table toolbar, there is one button that is particularly helpful in terms of sorting your samples. you can sort by file name, by sample name or severity, or the display name.

you can select the alleles that will display in both the cells and in the preview. you can choose allele, base pair, rfu, time, and peak area. you can also choose what to display in terms of file name, either the .fsa file name or the sample name in that file. when you open the graph view, you can do this a couple of different ways. there is a graph button at the upper left which opens the selected sample to the particular selected locus. you can double-click on a sample named open the sample, or you can double click on a cell to both open in the graph view and zoom to that particular locus. so this is what the graph view looks like if you've double clicked the file name. in the graph view, there is a toolbar as well that has a number of different buttons;

and we'll review these as we do the demonstration. when you hover over labels in either the preview or the graph view of electropherograms, you can hover over the allele labels or the artifact labels; and what will happen is you will get a box that opens up with information about what that peak has. in this case, this is an artifact. it says that the signal is a possible off-ladder allele and that the width of the peak is unusual. so at this point, i'd like to do a little bit of a demonstration of osiris -- the program itself. we'll do a live demonstration. i've installed this in the documents directory.

you can see that the logo pops up. in this case, because i've actually done some analysis using this installation, i have some recently viewed files -- close that window. i'm going to start a new analysis. in this particular case, the analysis window opens up. and what i need to do is i need to select the file that i want to analyze. you'll note that the osiris test analysis directory has a number of different sets of data that you can use as test data. i would recommend if you have data of your own that you try it with that.

but this is data that you can play around with. in particular, the identifier file folder has a folder that has a number of labeled artifacts in it. and if you want to see how it treats different artifacts, you can take a look at that. in this particular case today, we'll be looking at powerplex 16. this is sole source sample data from one of the nist mixture tests that they did. in this case, i've actually started another directory. i'm not putting my data into the osiris directory. i've made a subdirectory called osiris output. if you type a directory name in here, if it does not exist, osiris will create it for you.

then i need to select my operating procedure. you'll note that the majority of these are in brackets. the ones in brackets, these are predefined. they're defaults, and they can't be customized. you can use them as a template for your own custom analyses. so what i'm going to do is i'm going to use one that i set up for today specifically for this webinar. this is set up to model what a lot of casework laboratories are doing right now with relatively low analysis threshold of 50 rfu for the samples and 150 rfu for the ladders and 100 rfu for the ils. then i'm going to choose my ils.

i'm going to use the default one. you'll notice that this is actually all the same ils marker system; however, it uses more or fewer of the actual markers themselves. in this case, the highest molecular weight marker it uses is 500 base pairs. and you want to make sure that you have data for all of the markers selected in this ladder. some of these actually remove at least one of the markers that migrates anonymously. so i'll select the default and then i'll run it. because this is not very many samples, it takes a couple of seconds to run and then expands a full screen at this point.

it opens up in the table view. you'll notice that we have the three panels here. these panels can be modified to suit your screen better by dragging the edges. and you'll see that there is the standard colors set up here. so what we're going to look at is -- first of all, you can navigate by clicking any of these cells. you'll see that the white cells and the samples, like pause and control here, has passed because it has the green checkbox. some of these others have not passed because there are red x's. the loci that have issues -- for example, in this negative control, there is an artifact.

there is a peak here that's been called. and you can hover over the artifact label to see what that is. i'm expecting (inaudible) negative control. in here, you can navigate by clicking on the various cells; or you can use the mouse to navigate from cell to cell and from sample to sample. as you do that, you'll see the electropherogram show up in the preview pane. and you'll see down in the bottom pane, the notifications pane, you'll see that there are unexpected alleles in this particular case because we've got a third peak and this is expecting a sole source sample here.

in the left-hand panel, you'll see that there is peak information. it tells you all the data about the peak including the rfu; the time; the peak area; the fit, which is that measure of the quality of each (inaudible). so let's take a look at case no. 4 here. first of all, let's take a look at the sort porter. we actually sort these by either the file name, or we can sort by the severity. in this case, we're sorted by severity case no. 4 is this one here; and we can see that in this case, one of the sdr peaks was high in comparison to this. that's probably because of the heterozygous imbalance.

in this locus here, there is a peak here. one of the things you can do, if you right click, you can actually set various types of data in this particular view. so i think what we'll do now is we'll take a look at the stat graph here. so we can open either by clicking the graph button, or we can open it by double clicking the sample name itself, or we can open by double clicking the locus. i'm going to double click the locus. when we're opening the graph view, you'll see that this is pretty tight because of the screen size on the laptop that i'm using here for the presentation. what you should do is you should go to the graph menu and select resizable plots.

when you've done that, you can then grab the edge of the panes; and you can adjust the size so that you get an appropriate size for the screen that you're looking at. when you are in the graph view, there are a number of things you can do. you can zoom out, using the reset access button. you'll notice as i hold the cursor over that button that i get a tool tip that pops up which tells me what the button actually does. when you want to zoom in, you can actually do it by the classical click, drag and release, which will zoom in to where you want to go. if you want to zoom in, you can also use the keys -- "a" and the "d" key.

if you click in a particular pane in a particular electropherogram, you'll see that that will actually activate. you can see the little green dot there. once it's activated, you can use the "d" key to expand, to zoom in. you can use the "a" key to zoom back out. you can also use the "x" key to zoom the y axis out and the "w" key to zoom the y axis back in. in addition to that, and probably one of the easiest things to do, is you can zoom in by clicking any one of the locus markers, which will zoom you to that particular locus and zoom and scale you to it. in this particular case, we'll take a look at the different buttons on the toolbar.

the "a" button, which you can see here, is selected as analyze data. with any of these buttons, you can basically change all the plots at once by holding down the shift key when you do it. the "r" button will turn on the raw data. i'll turn the analyze data off. this is what the raw data looks like. before this is what osiris is actually using for its analysis. the "l" button will turn on the ladders. this gives you the ability to see how well the peak matches up with the ladder. you can see this has slightly shifted to the left.

using the ladder labels button, you can turn the ladder labels on and off depending on how busy you need it or whether you need that information. and finally, if you zoom out, this button here will display the baseline. so what i'm going to do is i'm going to take a look right here, and this display here is in light blue. it's the baseline that's been subtracted from the raw data. and as you can see, this stutter peak would be about half again as high in terms of its total rfu if the raw data hadn't been subtracted. if you're doing low-level analysis, if you're pushing your analysis threshold down below 75 -- or maybe even 100, depending on the quality of your data --

you should probably use baseline subtraction, which we call normalization. the multiple button will basically set all the panels to be the same in terms of what you've got displayed. the table button will let you swap back and forth between the table and the display. when you have an issue, such as this in d13, where this looks like it might be pull-up, what we can do is we can use the channel buttons to display, to overlay the different colors to determine whether or not it is in fact (inaudible). particularly when you look at the raw data, you can see that there is a little bit of a sort of s curve in there, which is pretty typical for pull-up peaks when you're looking at the raw data. in addition to that, there is the sync button.

if you unclick the sync button, that allows you to unsynchronize the various different axes, which is handy if you want to zoom in on a particular y axis. the ils button will display the ils marker in each of the channels. and the rfu button will display the rfu minimum analysis thresholds. when you're displaying labels, you can display -- change your labels on your peaks to base pairs, to your rfu value. and if you're looking for a display for (inaudible), you can actually display no labels at all. this is the artifact display. right now it's set for critical artifacts. so you'll see that this says that the base pair residual exceeds the threshold.

in other words, this has shifted beyond the threshold set by the laboratory. and this one, the stutter peak displays no label. if i change it to "all artifacts," that displays both the critical artifacts, such as this which requires review. it also displays non-critical artifacts, such as this stutter peak here. this tool tip box says this peak has restricted priority, and this is because of stutter. once you've got your data, you probably want to export it; and what i'm going to show you now is how to do that. i'm going to go back to the table. and first of all, i'm going to set up the export. when you try to do your first export,

if you look for it on the file menu you'll see that there is no exportive data aside from the export cmf file. that's because i haven't actually set one up yet. there are a couple of exports that we actually distribute with osiris. we also have a significant number of others. so that if you have a format requirement that's not being met, you can talk to us and we may have a solution for you already. okay, there are no export file types available. i'm going to select new to make a new one. in the config folder of the osiris install, there is the tab.

and this is described, by the way, in the osiris manual; you can set this up yourself using the users guide. i'm going to open that. it has chosen the default name of spreadsheet, which i'm going to leave. you can put your own name on it though if you like. actually, i'm going to go back. allow user to override, rule out or allow these or change the file name extension. you can just say next to that. this tip box right here will actually allow you to automatically run an export as soon as the analysis is completed, and that is without any other analyst intervention.

this allows laboratories to automate processes. you say, finish. this is my export; i will say, done. then when i go to file, you'll see that now it has an export. if you have multiples, it will give you an extra little dropdown to allow you to select different exports. i'll say, export spreadsheet. osiris then notifies me that i have not edited or dealt with four of these samples. you can see the four samples with x's. i'll say, yes, i want to explore the data anyway.

it allows me to choose a file name to do this. i will call it web export or export, as this case may be. and i will call this a tab file, and we'll save it. at that point, it asks you if you want the off-ladder allele labels in the table or not. i'll say, yes, because that gives me an indication of which ones are off-ladder alleles. most people, if you are importing this data, you don't want that because your lims system won't be happy with you if you do. if you click the view file location, that will actually allow you to open up the folder that has the file in it, which makes it easy to find and review.

show channel numbers in the column headings just basically puts a channel number in addition to the locus; most labs don't use that. i say, okay; and it opens it up. i'm going to look at that. here it is, at the top. i can then open that with various different files. in this case, i will choose -- i guess i won't choose -- see excel. here is excel. i can open it up, hit open, and there is your table in excel.

now obviously, you can set this up in almost any format. it's very, very flexible. you can set it up for any layout and any kind of data format as well. if you have particular requirements for your lims system or for programs that you use, feel free to ask us. but if you have somebody who is handy with xslt, you can program your own. okay, we'll close this and go back to our table. what i'd like to do now is briefly show you the lab settings -- how to set these up. you go to tools/lab settings, and there are keystrokes for all of these. the lab settings allow the users to set things up the way they want to, according to their own protocols.

there are various tabs here. the general tab allows you to choose between .fsa and .hid formatted data and to choose the default standard for this particular operating procedure. what you're going to need to do to start it is you're going to want to choose the template. in this case, let's choose a powerplex 16 template. this can't be edited because it's a default; so let's say, new, which is powerplex 16. we'll type in test. and then we can actually edit these things; we can edit and change the different things. so .fsa and .hid file formats, you may want to allow the user to override.

that allows the user to override the default internal marker, and it also allows the user to override -- actually, just the default marker there. you can type notes in about that particular operating procedure. you can also put protocol names and numbers here, lot numbers if you wish. for the files and sample names, you can define the strings in the file names or in the sample names; for ladder; for positive control/negative control for a possible mixture for a single-source sample; and then all of the other codis categories. what we recommend that you do is that you change the strings that you use for (inaudible) strings to identify these different sample to what your laboratory uses

and then delete the defaults that we've put in so that you won't find a sample that contains -al being identified as a ladder. you can also set the search to determine which of these samples are which. so in other words, when it's searching for ladders in this case, it's searching the file names for al and for the word ladder. but you can also set it to search within the sample names. the locus and ils thresholds tab allows you to choose thresholds for various filters, including a fractional filter, which is a global filter to remove samples below a certain percentage of highest peak within a locus. the same thing for pull-up peaks.

there is stutter, plus stutter, adenylation threshold for minus adenylation; heterozygote balance and homozygote threshold, which will allow you to determine what level homozygote peaks are too low to pass automatically -- aren't tall enough. for sample thresholds, you can do the analysis thresholds and other various thresholds here. and you can set a variety of different parameters, which are described in the users guide. i won't go through them all, but you can certainly look them up in the users guide. the ones that i would point out, however, are normalized raw data relative to baseline. that is one you that you should set if you're using -- that will do the baseline subtraction if you are doing low-level thresholds analysis, if you're trying to push your thresholds below 75.

and you should take a look at the disable low level height filters from the mixtures. that is described in the users guide, but that will allow you to differentially apply these various filters -- fractional filter, stutter filter, adenylation filter -- so that you can actually analyze both reference samples and mixture samples in the same analysis and apply filters only to the reference samples, the known samples, so that you can then put that data directly into a mixture calculation software. a number of different laboratories are using osiris to feed their data into their mixture software. finally, there is the assignments tab. the off-ladder alleles allows you to predefine off-ladder alleles so that the program will recognize them

as being accepted alleles and will not complain about them. you won't get artifact notifications. there is also a tab that allows you to define your own custom positive controls so that the program will actually check the correct allele calls within those custom-positive controls or to add new kit-positive controls to the kit definition in your operating procedure. and finally, the accept and review tab, which allows you to set the number of reviews that is required by the laboratory, which osiris will then enforce in the software. if you are evaluating this, you want to check this here: allow reviewer to modify user name.

that allows an evaluator to serve both as the first reviewer and as the second reviewer without having to log in under a different user name. and then you just say, okay. one of the things that is worthwhile noting is when you're actually modifying the user settings, sometimes the lock button at the bottom left will not be grayed out. and you will not be able to change your settings. and that is to prevent you from changing the settings while analysis is being run or for another user when multiple users are using it on the network to prevent them from changing the settings while you are actually running your analysis.

if the lock button is lit up, then what you can do is click the lock button until it greys back out; and then you'll be able to modify the settings. one of the buttons in the graph view that is worth noting is the parameters button. the parameters button is particularly handy because it gives you both the name of the input directory and a link. if you click it, it will actually open the directory up -- the input directory and the output directory and the operating procedure name as well. be aware that if you click the parameters button, it is not the same thing as if you open the tools lab settings button because if you open the parameters here -- the lab settings from these parameters -- you'll see that everything is grayed out.

this is a historical record of what was used to run this particular analysis. it is stored in the analysis file itself and can't be modified. you're going to have to go back and change the lab settings using the tools menu and rerun your analysis if you want to change settings here. the other thing that's handy here is if you open the parameters from the graph view, it will tell you which ladder file was used to do the analysis in case the ladder file itself has artifacts in it. generally, osiris is very robust in analyzing ladders; it rarely makes mistakes. but sometimes it does point out artifacts, such as spikes, that are not problematical in terms of the correctness of the ladder analysis;

but we want to make sure that we point those artifacts out. at this point, i think that we will probably open this up for questions. if you have any questions and you haven't typed them in yet, please go ahead and do that. we will try to address some of these questions as we go along. the first question i see is: is osiris used for semi-validation? and the answer is, yes. there (inaudible). we would be happy to discuss settings for semi validation separately. these settings are really not intended for a set semi validation but you can see if we set it up for that,

in fact i talked to you about that. is the keyprint 10 system (inaudible)? that can certainly be added. you can add kits to osiris. there is a description of how to do that in the user guide, and we would be happy to help you out with that. if there are commercial kits that are not in osiris and somebody wants to add one of those, we would be happy once the kit is defined to add that to the distribution for future releases. how do you (inaudible)? i'm not sure exactly what (inaudible).

if the questioner would expand a little bit more, we'll see if we can come back to that. do you have the target mass of the powerplex 16 mixer? it's not clear quite what you mean by the target mass. but certainly the area under the curve, yes, that is available for each individual peak, not necessarily in total for the entire curve. osiris significance to a student not into forensics? this is actually a tremendously helpful program for training if you're looking at str analysis because one of the things that it does is it tells you exactly what it thinks each peak's problem is if it's an artifact.

so it will tell you whether it thinks it is pull-up. it will tell you if it's below scale, if it's stutter, if it's adenylation. so it can be very, very helpful in terms of student training. it's been used to do training for forensic users. can you click off interfacts directly from the (inaudible). this is on our to-do list, and this is in the development planned for the immediate future -- not yet, but very soon. and again, i would encourage people to sign up on the announcements list so that you get notified of the latest and greatest releases. that is something that (inaudible) to make it much easier much sooner.

notification of video release? yes, we will notify the people who signed up for the webinar when the video is released and also as to exactly how to do that. our next question is: can i view the sample information, the injection time and the voltage? at this point, you can't do that in osiris; but that is something that we are actually also looking at that's a user request for development. and that's definitely something that we have on our to-do list. next question is: how do you estimate the threshold value? every laboratory estimates their own.

the forensic laboratories are typically nowadays beginning to do this by looking at the noise levels within the channels for their particular system. every laboratory is going to have differences because of the different machinery and just the way it's set up. laboratories have used everything from 150 rfu for .fsa files all the way down to below 50 rfu. if you're using .hid files, you'll probably -- because they have a much higher signal level, they also have a much higher noise level -- you'd have to choose a higher threshold value. to some extent, if you don't want to look at the noise -- if you don't want to do the noise calculations -- you can actually just do this interim, just try it out. can you manually edit allele calls?

yes, you can definitely manually edit allele calls. you can actually change the number, and we can probably do a quick demonstration of that. possibility of support for startup use? we do have --there is an e-mail address that you can send questions to, and we definitely try to help our users as much as we can. we have some training materials on its use, which we can make available -- eventually we'll make available on the web. for non-human strs? sure, this can be used for non-human strs. that would mean putting a kit in.

you'd need to have a ladder, but you could make a ladder yourself. it wouldn't be that difficult to do, and i'd be happy to discuss that with anybody. please send us an e-mail. the e-mail address is: forensics@ncbi.nlm.nrh.gov. and we'll actually put that up and e-mail that to the users in addition. can you override the ladder requirement? i'm not sure exactly what that question means. you need to have a ladder, but there are possibilities for using non-traditional ladders which we'd be happy to discuss with somebody.

to create new panels of bins? that would be creating a new kit definition. that's described in the users manual; and again, if you have any issues, if you have any questions, we'd be happy to work with you on those. i want to do a really quick little editing demonstration. we'll take a look at our simplest case right here. okay, so let me show you editing and allele call since we have a few minutes left. so i've selected this allele call here. this, it turns out if you take a look, is probably a spike right in here.

you'll notice that there is an allele, an artifact notification. we'll go back, and we will edit that call off. what you do is you go to the table menu. there are a number of ways of getting at this, but i'm going to show you the simplest way. edit d13. then the edit box comes up; and at this point what i can do is -- this is peak 1 is the nine-peak right here that i want to get rid of. i'm going to simply untick the allele peak box. i'm going to say, 9 is spike; so that i'm going to create an audit trail

which will show up in the review, acceptance and notes here. i'll say, yes; you'll notice that the allele call is gone. you could also go back and i could change that to add the artifact peak notice as well because i want to make sure that that shows up. you'll see now that my notes are showing up here. that remains as part of the permanent record in the file. and i'm going to say (inaudible) and say, okay. and then when i look at that -- it's not showing up. we'll take a look at that and see.

oh, that's because i have critical set. thank you, doug; that's correct. when i show all artifacts, that will come back up. you'll see if i box something and don't want to blow up to a tiny little pixel, i can actually move back and it will not blow up. so, other questions -- is there an audit trail available? the audit trail is in the software. if you wanted an audit trail or a printable audit trail, that's something that we could certainly examine in terms of development.

but currently, you can't print it out separately. a number of laboratories are looking to go paperless using osiris and the files themselves as the documentation for what they've done. is it possible to add more than one label for a plot? the answer to that is not yet, but very soon. that's actually been developed by some collaborators at (inaudible) state university, and that is something that we're going to put in -- although possibly not this next release which is out very soon, but almost certainly in the release following that -- which shouldn't be that far off. is it possible to print the electropherogram?

actually, i think what we'll do is we'll follow up with additional questions via e-mail to folks because it's about time to wrap up the meeting. ben, do you have anything to add? we can also post these on faq. we will also post answers to these questions on faq. i'd like to thank you all for attending the webinar. if you have any questions, again, we will be happy to answer them in the faq for the ones you typed in. you can send your questions to the e-mail, which is on the osiris home page. and that goes into our queue, and we will get back to you on that as well.