29 May 2023

Poll results: fragment libraries in 2023

Our latest poll on fragment libraries suggests the field is settling into some standard practices. The poll ran from April 9 through May 26. Of the 59 participants, all but one answered all the questions (there was one skip for the last question). This is slightly down from previous years; perhaps people are sick of internet polls? We also don’t know how many organizations the respondents represent; it is possible several people voted from one company or university, which might skew the results. Nonetheless, we think this survey gives a reasonable snapshot of how people construct and maintain fragment libraries.
Our first question asked about library size, and the results are similar to when we last asked this question in 2018, with the average library having between 1001 and 2000 fragments.
Next, we asked about the size of fragments themselves, specifically the minimum and maximum number of non-hydrogen atoms allowed in a fragment. The minimum hasn’t really changed from 2018, averaging 7-8 heavy atoms. However, the fraction of respondents who include the tiniest fragments has doubled (albeit from a low number), perhaps due to increasing interest in MiniFrags and MicroFrags.
Unlike in 2018, the maximum size of fragments seems to be bimodal, with some folks drawing the line at 15-16 heavy atoms (consistent with this analysis from Astex) while others allow larger fragments. It will be interesting to see whether this bifurcation represents a true shift, though even fragments with 22 heavy atoms are likely to be under 300 Da, consistent with the rule of three, which is twenty years old this year.
We then asked about the presence of chiral molecules. There was little change from 2017, with most respondents stating that they have racemic compounds in their library, though there was a slight increase in the number of respondents excluding chiral fragments.
A new question for this poll asked whether synthetic tractability was considered at the outset of library design. This was a consideration for 85% of people who took the poll; more than a third said they considered progressability for every fragment in the library.

Finally, we asked about library storage conditions. As was the case when we asked this question nine years ago, more than two-fifths of respondents said they store their library at -20 ˚C. However, the fraction of respondents who store their library at room temperature dropped, while those who store their libraries at -80 ˚C increased.
Although some changes are noticeable over the years, it seems that best practices have been established and widely adopted in fragment library design. What do you think – does anything surprise you?