Fall 2024 Feature Releases 🍂

Normalization Baseline Selection Feature For Data Export

We’ve added a new feature in the data export tab, enabling you to select a normalization baseline for your data. Now, you can choose any time point as the baseline, and the data will be normalized accordingly. The default option is set to “no normalization,” allowing for complete flexibility based on your analysis needs.

Enhanced Navigation with New Collapsible Filter

We’ve upgraded our filtering interface with a new collapsible design. This accordion-style filter allows for easier navigation and quicker access to the metadata you need, streamlining the process of selecting filters and viewing data.

The previous layout required extensive scrolling to reach different filter options. With the collapsible filter, all options are now neatly organized and accessible, enabling you to focus on analysis without distractions.

Append: add new samples to your web app as the samples get processed

Customers who send us samples in rolling cohorts can see their samples updated as new samples come in. Imagine you have a 250 specimen trial. 100 specimens arrive in Cohort 1, and 150 specimens arrive in Cohort 2.

Previously, the addition of new samples led to a short period of downtime for the user’s web app as the new samples were added to the project. Now, samples are added to the web app without any downtime, and all statistics are recomputed for all samples. This update provides a more seamless experience for the customer.

“No samples left behind”: Missing a baseline sample, no sweat!

Previously, if a sample was missing in a certain group comparison, that group comparison would not yield statistical results. For example, assume you have a set of five patients.

If you want to compare B cells across two on-treatment timepoints (i.e. C1D2 and C1D15) normalized to the baseline, but one subject is missing the baseline sample, previously our statistical testing implementation would skip the comparison due to the incomplete data.

Let’s see this in action. If Patient two in the table below were missing a baseline sample, we wouldn’t be able to calculate statistical results. Now, by dynamically constructing the statistical test to remove the one patient with no baseline sample, we are able to calculate all statistical tests for the remaining four samples.

Patient #	Baseline	C1D2	C1D15	C2D1	C2D15
1	✅	✅	✅	✅	✅
2		✅	✅	✅	✅
3	✅	✅	✅	✅	✅
4	✅	✅	✅	✅	✅
5	✅	✅	✅	✅	✅

Now, say a subject is missing any timepoint (as in the table below). And assume the goal is to calculate paired statistics. Previously, if you tried comparing the means of these two groups at two different timepoints, no statistics would be shown. Our algorithm was strict and simply did not perform the comparison. Now, we filter out the subject with the missing sample and show statistics for the remaining patients.

Patient #	C1D2	C1D15
1	✅	✅
2	✅
3	✅	✅
4	✅	✅
5	✅	✅