BLEND

J Foadi, P Aller, G Evans

BLEND is a program that aims to accelerate and assist in automation of the analysis of partial diffraction data sets measured from multiple crystals samples. This problem is very common nowadays where the use of small crystals and micro-beams results in numerous partial data sets all suffering to varying degrees from radiation damage.

BLEND uses cluster analysis techniques prior to the scaling and merging of integrated data sets to categorise data sets into groups that will be more likely to scale and merge well together. Our aims are

  • allow the user to carry forward multiple scaling groups that do not necessarily scale together but may scale well with other data sets still to be measured
  • in doing so make more efficient use of all data measured by avoiding the practice of rejecting large numbers of outliers at the scaling stage
  • create groups of data that will produce more stable scaling jobs thereby easing the job of automating these tasks
  • accelerate the process of merging together many partial data sets by guiding the user at to the best possible combinations of data
  • BLEND generates multiple scaling groups of data and as its output offers the user a choice of results based on the compromise of final data set completeness versus quality of merging statistics.

BLEND is available within the latest CCP4 release or from the authors on request.