Back

Vasculature-staining with lipophilic dyes in tissue-cleared brains assessed by deep learning

Henriksen, B. L. E.; Jensen, K. H. R.; Berg, R. W.

2020-05-17 neuroscience
10.1101/2020.05.16.099705 bioRxiv
Show abstract

Visualization of the vasculature in three dimensions (3D) has become attractive, particularly in stroke models. 3D-reconstruction is aided by tissue-clearing, where the transparency allows imaging of fluorescent probes in deeper structures. The vasculature is commonly stained by fluorescent lipophilic dyes that are incorporated into the wall during transcardial perfusion. Nevertheless, tissue clearing involves extracting the light-scattering lipids, and hence also the lipid-appended dyes. The wash-out likely depends on dye and its aldehyde-fixability. Fixation secures cross-linking to proteins and hence retainment in the tissue. However, the compatibility of various types of dyes is largely unknown. We tested and compared 9 different dyes for vasculature staining and tolerance to lipid clearing, which was quantified using deep learning image segmentation. Among the dyes, we found a subset that is both cost-effective and compatible with tissue lipid clearing. We suggest these dyes will provide a valuable tool for future investigations. HighlightsO_LISize, alkyl-chains and aldehyde-fixable groups improve dye retention and performance. C_LIO_LICost-effective dyes in specific liposomes result in optimal vessel staining. C_LIO_LISP-DiIC18 is compatible with CLARITY and BrainFilm and advantageous for most studies. C_LIO_LIWe recommend, SP-DiI and R18, cost-effective dyes, for vessel-painting with CLARITY. C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=103 SRC="FIGDIR/small/099705v1_ufig1.gif" ALT="Figure 1"> View larger version (30K): org.highwire.dtl.DTLVardef@2f7eborg.highwire.dtl.DTLVardef@e272a4org.highwire.dtl.DTLVardef@d48a29org.highwire.dtl.DTLVardef@1ba9781_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 11 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.